WO2020155142A1 - 一种点云重采样的方法、装置和系统 - Google Patents

一种点云重采样的方法、装置和系统 Download PDF

Info

Publication number
WO2020155142A1
WO2020155142A1 PCT/CN2019/074585 CN2019074585W WO2020155142A1 WO 2020155142 A1 WO2020155142 A1 WO 2020155142A1 CN 2019074585 W CN2019074585 W CN 2019074585W WO 2020155142 A1 WO2020155142 A1 WO 2020155142A1
Authority
WO
WIPO (PCT)
Prior art keywords
point cloud
grid
initial point
resampling
module
Prior art date
Application number
PCT/CN2019/074585
Other languages
English (en)
French (fr)
Inventor
李延召
张富
洪小平
Original Assignee
深圳市大疆创新科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳市大疆创新科技有限公司 filed Critical 深圳市大疆创新科技有限公司
Priority to CN201980005631.3A priority Critical patent/CN111801707A/zh
Priority to PCT/CN2019/074585 priority patent/WO2020155142A1/zh
Publication of WO2020155142A1 publication Critical patent/WO2020155142A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/50—Depth or shape recovery

Definitions

  • the present invention generally relates to the technical field of distance measuring devices, and more specifically to a point cloud resampling method, device and system.
  • Lidar is a perception system of the outside world. By transmitting and receiving and processing light wave information, it can learn the three-dimensional information of the outside world. It is no longer limited to the plane perception of the outside world such as cameras.
  • the uniformity and noise of the point cloud output by the laser radar ranging device are important parameters of the point cloud quality, which affect the display effect and further processing of the point cloud.
  • improving the quality of point clouds is usually done by improving hardware sampling and point cloud generation algorithms.
  • the above-mentioned solutions have higher requirements for the entire system, and usually ignore the other and are restricted by the difficulty of system construction, power consumption, cost, and scanning rate. In practice, the point cloud output by the optimized design still has some quality problems.
  • one aspect of the present invention provides a point cloud resampling method, the method includes:
  • the distribution uniformity is higher than the distribution uniformity of the initial point cloud, and/or the noise of the resampled point cloud is lower than the noise of the initial point cloud.
  • the re-sampling of the initial point cloud output by the distance measuring device includes:
  • Interpolation is performed at the point cloud gap of the initial point cloud.
  • the projection points of the resampled point cloud on the reference surface have a higher uniformity than the projection points of the initial point cloud on the reference surface.
  • the reference surface includes at least one of a flat surface, a spherical surface, and a cylindrical surface.
  • the plane is a plane perpendicular to the central axis of the light pulse sequence emitted by the distance measuring device.
  • the spherical surface is a spherical surface centered on the distance measuring device; or,
  • the central axis of the cylindrical surface is a vertical line passing through the distance measuring device.
  • At least part of the projection points of the resampled point cloud on the reference surface are uniformly distributed.
  • At least some of the projection points of the resampled point cloud on the reference surface are arranged at equal intervals or at equal angles on at least one track of the reference surface.
  • the trajectory includes a trajectory in a regular straight line, a regular curve, or a regular broken line.
  • the track includes a track extending in a horizontal direction and/or a vertical direction.
  • the trajectory includes a plurality of concentric rings.
  • the centers of the plurality of concentric rings are located on the central axis of the distance measuring device.
  • the track includes a spiral track.
  • the trajectory includes multiple trajectories radiating outward along a central point.
  • the resampled point cloud adds a plurality of point cloud points located on the same reference surface.
  • the interpolation method includes: projecting the initial point cloud onto a reference surface to obtain projection points of the initial point cloud;
  • An interpolation point is determined at the gap between the projection points on the reference surface to obtain an image on the reference surface.
  • the depth value of the interpolation point is determined based on the depth value of the adjacent projection point.
  • the interpolation method further includes: discretizing the projection point of the initial point cloud on the reference surface before determining the interpolation point.
  • the method of discretization includes:
  • the projection points of the initial point cloud are gridded on the reference surface, wherein the positions of the interpolation points include at least partially empty grids.
  • the re-sampling of the initial point cloud acquired by the distance measuring device includes:
  • Resampling is performed based on the image to obtain the resampled point cloud with a specific arrangement pattern.
  • the method of gridding includes:
  • the depth value of the interpolation point is determined based on the depth value of the adjacent grid.
  • the method for determining the interpolation point further includes:
  • the position of the empty grid is a sky position, wherein the empty grid to be interpolated includes an empty grid of a non-sky position.
  • the grid is formed by a plurality of mutually perpendicular tracks on the reference surface.
  • the grid is formed by multiple crossing tracks extending in the horizontal direction and multiple crossing tracks extending in the vertical direction on the reference plane.
  • the grid is formed by intersecting multiple concentric circles on the reference surface and multiple tracks radiating outward from the center of the concentric rings.
  • the grid is formed by a spiral line on a reference surface intersecting multiple tracks radiating from the center of the spiral line.
  • the method further includes:
  • a certain interpolation point is added to the initial point cloud.
  • different areas of the initial point cloud have different point cloud densities
  • re-sampling the initial point cloud includes:
  • the scanning density of the distance measuring device in the central area of the scanning field of view is higher than the scanning density of other areas
  • the re-sampling of the initial point cloud includes: down-sampling the initial point cloud in the central area.
  • the method of downsampling includes:
  • the method of downsampling includes:
  • the method of restricting the number of initial point cloud points falling below a threshold number in at least part of the grid in the central area includes:
  • re-sampling the initial point cloud includes: performing noise reduction and re-sampling on the initial point cloud, so that the noise of the re-sampled point cloud is lower than the noise of the initial point cloud.
  • the method for noise reduction and resampling includes:
  • the method for noise reduction and resampling further includes:
  • each initial point cloud point is adjusted to the depth value of its corresponding pixel, or adjusted to the depth value of the adjacent pixel of the initial point cloud point at the projection position of the image.
  • the method for noise reduction and resampling further includes:
  • the image includes a depth distribution map.
  • the method of obtaining the image includes:
  • the method of obtaining the image includes:
  • the depth value of the empty grid is determined based on the depth value of the neighboring grid.
  • the method for determining the depth value of the empty grid based on the depth value of the neighboring grid is specifically:
  • the distance measuring device includes a transmitting module for emitting a light pulse sequence, and a scanning module for changing the direction of the light pulse sequence to scan the field of view.
  • the scanning paths of the scanning modules are different at least at some different moments.
  • the scanning area of the scanning module in the field of view of the distance measuring device increases with the accumulation of time.
  • the scanning module includes at least one light refraction element having a non-parallel exit surface and an entrance surface.
  • the scanning module includes two or three light refraction elements sequentially arranged on the exit optical path of the light pulse sequence.
  • At least two of the light refraction elements in the scanning module rotate during the scanning process to change the direction of the light pulse sequence.
  • the point cloud resampling device includes: a resampling module, configured to resample the initial point cloud acquired by the distance measuring device to obtain the resampling after resampling Point cloud, wherein the distance measuring device has a non-uniform scanning density in the scanning field of view, the distribution uniformity of the resampled point cloud is higher than the distribution uniformity of the initial point cloud, and/or, The noise of the resampled point cloud is lower than the noise of the initial point cloud.
  • the point cloud resampling system includes:
  • At least one ranging device for detecting the target scene to generate an initial point cloud
  • Memory used to store executable instructions
  • the processor is configured to execute the instructions stored in the memory, so that the processor executes the aforementioned point cloud resampling method.
  • the initial point cloud with irregular sampling patterns can be converted into a uniformly sampled resampled point cloud, and the sampling density of the point cloud can be increased or decreased.
  • the resampled point cloud can be more suitable for subsequent processing, and/or have better display effects, and/or hide specific hardware sampling modes, etc., and the noise of the resampled point cloud is lower than the initial point cloud The noise, thus improving the accuracy of ranging.
  • FIG. 1 shows a schematic structural diagram of a distance measuring device in an embodiment of the present invention
  • Figure 2 shows a schematic diagram of a distance measuring device in an embodiment of the present invention
  • Fig. 3 shows a schematic diagram of an initial point cloud obtained by a distance measuring device in an embodiment of the present invention
  • FIG. 4 shows a schematic diagram of comparison between the initial point cloud and the resampled point cloud in an embodiment of the present invention
  • FIG. 5 shows a schematic diagram of point cloud projection in an embodiment of the present invention
  • Fig. 6 shows a schematic diagram of the projection of the initial point cloud on the reference plane in an embodiment of the present invention
  • Fig. 7 shows a point cloud projection diagram after discretization in an embodiment of the present invention
  • Figure 8 shows a depth distribution diagram after interpolation in an embodiment of the present invention
  • Fig. 9 shows a schematic diagram of a sampling coordinate system in an embodiment of the present invention.
  • FIG. 10 shows a schematic diagram of projection points of a resampled point cloud on a reference surface in an embodiment of the present invention
  • Figure 11 shows a schematic diagram of a resampled point cloud in an embodiment of the present invention
  • Fig. 12 shows a schematic diagram of projection points of a resampled point cloud on a reference surface in another embodiment of the present invention
  • Figure 13 shows a schematic diagram of a resampled point cloud in another embodiment of the present invention.
  • FIG. 14 shows a schematic diagram of projection points of a resampled point cloud on a reference surface in still another embodiment of the present invention.
  • FIG. 15 shows a schematic diagram of a resampled point cloud in still another embodiment of the present invention.
  • FIG. 16 shows a schematic diagram of projection points of a resampled point cloud on a reference surface in another embodiment of the present invention.
  • Figure 17 shows a schematic diagram of a resampled point cloud in another embodiment of the present invention.
  • FIG. 18 shows a schematic block diagram of a point cloud resampling device in an embodiment of the present invention
  • Fig. 19 shows a schematic block diagram of a point cloud resampling system in an embodiment of the present invention.
  • the distance measuring device includes a lidar.
  • the distance measuring device is only used as an example. Distance devices can also be applied to this application.
  • the distance measuring device may be electronic equipment such as lidar and laser distance measuring equipment.
  • the distance measuring device is used to sense external environmental information, for example, distance information, orientation information, reflection intensity information, speed information, etc. of environmental targets.
  • the distance measuring device can detect the distance from the probe to the distance measuring device by measuring the time of light propagation between the distance measuring device and the probe, that is, the time-of-flight (TOF).
  • the ranging device can also detect the distance from the detected object to the ranging device through other technologies, such as a ranging method based on phase shift measurement or a ranging method based on frequency shift measurement. This is not limited.
  • the distance measuring device includes a transmitting module, a receiving module, and a temperature control system.
  • the transmitting module is used to emit light pulses;
  • the receiving module is used to receive at least part of the light pulses reflected back by the object, and according to the received at least The partial light pulse determines the distance of the object relative to the distance measuring device.
  • the transmitting module includes a transmitting circuit 110; the receiving module includes a receiving circuit 120, a sampling circuit 130 and an arithmetic circuit 140.
  • the transmitting circuit 110 may emit a light pulse sequence (for example, a laser pulse sequence).
  • the receiving circuit 120 may receive the light pulse sequence reflected by the object to be detected, and perform photoelectric conversion on the light pulse sequence to obtain an electrical signal. After processing the electrical signal, it may be output to the sampling circuit 130.
  • the sampling circuit 130 can sample the electrical signal to obtain the sampling result.
  • the arithmetic circuit 140 may determine the distance between the distance measuring device 100 and the detected object based on the sampling result of the sampling circuit 130.
  • the distance measuring device 100 may further include a control circuit 150, which can control other circuits, for example, can control the working time of each circuit and/or set parameters for each circuit.
  • a control circuit 150 can control other circuits, for example, can control the working time of each circuit and/or set parameters for each circuit.
  • the distance measuring device shown in FIG. 1 includes a transmitting circuit, a receiving circuit, a sampling circuit, and an arithmetic circuit for emitting a light beam for detection
  • the embodiment of the present application is not limited to this, the transmitting circuit
  • the number of any one of the receiving circuit, the sampling circuit, and the arithmetic circuit can also be at least two, which are used to emit at least two light beams in the same direction or in different directions; wherein, the at least two light paths can be simultaneous Shooting, or shooting at different times.
  • the light-emitting chips in the at least two transmitting circuits are packaged in the same module.
  • each emitting circuit includes a laser emitting chip, and the dies in the laser emitting chips in the at least two emitting circuits are packaged together and housed in the same packaging space.
  • the distance measuring device 100 may also include a scanning module for changing the propagation direction of at least one light pulse sequence (for example, a laser pulse sequence) emitted by the transmitting circuit, so as to control the field of view.
  • a scanning module for changing the propagation direction of at least one light pulse sequence (for example, a laser pulse sequence) emitted by the transmitting circuit, so as to control the field of view.
  • the scanning area of the scanning module in the field of view of the distance measuring device increases with the accumulation of time.
  • the module including the transmitting circuit 110, the receiving circuit 120, the sampling circuit 130, and the arithmetic circuit 140, or the module including the transmitting circuit 110, the receiving circuit 120, the sampling circuit 130, the arithmetic circuit 140, and the control circuit 150 may be referred to as the tester.
  • Distance module the distance measurement module can be independent of other modules, for example, scanning module.
  • a coaxial optical path can be used in the distance measuring device, that is, the light beam emitted by the distance measuring device and the reflected light beam share at least part of the optical path in the distance measuring device.
  • the distance measuring device can also adopt an off-axis optical path, that is, the light beam emitted by the distance measuring device and the reflected light beam are transmitted along different light paths in the distance measuring device.
  • Fig. 2 shows a schematic diagram of an embodiment in which the distance measuring device of the present invention adopts a coaxial optical path.
  • the ranging device 200 includes a ranging module 210, which includes a transmitter 203 (which may include the above-mentioned transmitting circuit), a collimating element 204, a detector 205 (which may include the above-mentioned receiving circuit, sampling circuit, and arithmetic circuit) and Light path changing element 206.
  • the ranging module 210 is used to emit a light beam, receive the return light, and convert the return light into an electrical signal.
  • the transmitter 203 can be used to transmit a light pulse sequence.
  • the transmitter 203 may emit a sequence of laser pulses.
  • the laser beam emitted by the transmitter 203 is a narrow-bandwidth beam with a wavelength outside the visible light range.
  • the collimating element 204 is arranged on the exit light path of the emitter, and is used to collimate the light beam emitted from the emitter 203, and collimate the light beam emitted from the emitter 203 into parallel light and output to the scanning module.
  • the collimating element is also used to condense at least a part of the return light reflected by the probe.
  • the collimating element 204 may be a collimating lens or other elements capable of collimating light beams.
  • the transmitting light path and the receiving light path in the distance measuring device are combined before the collimating element 204 through the light path changing element 206, so that the transmitting light path and the receiving light path can share the same collimating element, so that the light path More compact.
  • the transmitter 203 and the detector 205 may use respective collimating elements, and the optical path changing element 206 may be arranged on the optical path behind the collimating element.
  • the optical path changing element can use a small area mirror to The transmitting light path and the receiving light path are combined.
  • the optical path changing element may also use a reflector with a through hole, where the through hole is used to transmit the emitted light of the emitter 203 and the reflector is used to reflect the returned light to the detector 205. In this way, the shielding of the back light by the bracket of the small mirror in the case of using the small mirror can be reduced.
  • the optical path changing element deviates from the optical axis of the collimating element 204.
  • the optical path changing element may also be located on the optical axis of the collimating element 204.
  • the distance measuring device 200 further includes a scanning module 202.
  • the scanning module 202 is placed on the exit light path of the distance measuring module 210.
  • the scanning module 202 is used to change the transmission direction of the collimated light beam 219 emitted by the collimating element 204 and project it to the external environment, and project the return light to the collimating element 204 .
  • the returned light is collected on the detector 205 via the collimating element 204.
  • the scanning module 202 may include at least one optical element for changing the propagation path of the light beam, wherein the optical element may change the propagation path of the light beam by reflecting, refraction, or diffracting the light beam, for example,
  • the optical element includes at least one light refraction element having a non-parallel exit surface and an entrance surface.
  • the scanning module 202 includes a lens, a mirror, a prism, a galvanometer, a grating, a liquid crystal, an optical phased array (Optical Phased Array), or any combination of the foregoing optical elements.
  • at least part of the optical elements are moving.
  • a driving module is used to drive the at least part of the optical elements to move.
  • the moving optical elements can reflect, refract, or diffract the light beam to different directions at different times.
  • the multiple optical elements of the scanning module 202 can rotate or vibrate around a common axis 209, and each rotating or vibrating optical element is used to continuously change the propagation direction of the incident light beam.
  • the multiple optical elements of the scanning module 202 may rotate at different speeds or vibrate at different speeds.
  • at least part of the optical elements of the scanning module 202 may rotate at substantially the same rotation speed.
  • the multiple optical elements of the scanning module may also be rotated around different axes.
  • the multiple optical elements of the scanning module may also rotate in the same direction or in different directions; or vibrate in the same direction, or vibrate in different directions, which is not limited herein.
  • the scanning module 202 includes a first optical element 214 and a driver 216 connected to the first optical element 214.
  • the driver 216 is used to drive the first optical element 214 to rotate around the rotation axis 209 to change the first optical element 214.
  • the direction of the beam 219 is collimated.
  • the first optical element 214 projects the collimated beam 219 to different directions.
  • the angle between the direction of the collimated beam 219 changed by the first optical element and the rotation axis 209 changes as the first optical element 214 rotates.
  • the first optical element 214 includes a pair of opposed non-parallel surfaces through which the collimated light beam 219 passes.
  • the first optical element 214 includes a prism whose thickness varies in at least one radial direction.
  • the first optical element 214 includes a wedge-angle prism, and the collimated beam 219 is refracted.
  • the scanning module 202 further includes a second optical element 215, the second optical element 215 rotates around the rotation axis 209, and the rotation speed of the second optical element 215 is different from the rotation speed of the first optical element 214.
  • the second optical element 215 is used to change the direction of the light beam projected by the first optical element 214.
  • the second optical element 215 is connected to another driver 217, and the driver 217 drives the second optical element 215 to rotate.
  • the first optical element 214 and the second optical element 215 can be driven by the same or different drivers, so that the rotation speed and/or rotation of the first optical element 214 and the second optical element 215 are different, so as to project the collimated light beam 219 to the outside space.
  • the controller 218 controls the drivers 216 and 217 to drive the first optical element 214 and the second optical element 215, respectively.
  • the rotational speeds of the first optical element 214 and the second optical element 215 can be determined according to the expected scanning area and pattern in actual applications.
  • the drivers 216 and 217 may include motors or other drivers.
  • the second optical element 215 includes a pair of opposite non-parallel surfaces through which the light beam passes. In one embodiment, the second optical element 215 includes a prism whose thickness varies in at least one radial direction. In one embodiment, the second optical element 215 includes a wedge prism.
  • the scanning module 202 further includes a third optical element (not shown) and a driver for driving the third optical element to move.
  • the third optical element includes a pair of opposite non-parallel surfaces, and the light beam passes through the pair of surfaces.
  • the third optical element includes a prism whose thickness varies in at least one radial direction.
  • the third optical element includes a wedge prism. At least two of the first, second, and third optical elements rotate at different rotation speeds and/or rotation directions.
  • the scanning module includes two or three light refraction elements arranged in sequence on the exit light path of the light pulse sequence.
  • at least two of the light refraction elements in the scanning module rotate during the scanning process to change the direction of the light pulse sequence.
  • the scanning path of the scanning module is different at least partly at different moments.
  • the rotation of each optical element in the scanning module 202 can project light to different directions, such as the direction of the projected light 211 and the direction 213, so that the distance measurement device 200 Space to scan.
  • the light 211 projected by the scanning module 202 hits the detection object 201, a part of the light is reflected by the detection object 201 to the distance measuring device 200 in a direction opposite to the projected light 211.
  • the return light 212 reflected by the probe 201 is incident on the collimating element 204 after passing through the scanning module 202.
  • the detector 205 and the transmitter 203 are placed on the same side of the collimating element 204, and the detector 205 is used to convert at least part of the return light passing through the collimating element 204 into an electrical signal.
  • an anti-reflection film is plated on each optical element.
  • the thickness of the antireflection film is equal to or close to the wavelength of the light beam emitted by the emitter 203, which can increase the intensity of the transmitted light beam.
  • a filter layer is plated on the surface of an element located on the beam propagation path in the distance measuring device, or a filter is provided on the beam propagation path for transmitting at least the wavelength band of the beam emitted by the transmitter, Reflect other bands to reduce the noise caused by ambient light to the receiver.
  • the transmitter 203 may include a laser diode through which nanosecond laser pulses are emitted.
  • the laser pulse receiving time can be determined, for example, the laser pulse receiving time can be determined by detecting the rising edge time and/or the falling edge time of the electrical signal pulse.
  • the distance measuring device 200 can calculate the TOF using the pulse receiving time information and the pulse sending time information, so as to determine the distance between the probe 201 and the distance measuring device 200.
  • the distance and orientation detected by the distance measuring device 200 can be used for remote sensing, obstacle avoidance, surveying and mapping, modeling, navigation, and the like.
  • the distance measuring device of the embodiment of the present invention can be applied to a mobile platform, and the distance measuring device can be installed on the platform body of the mobile platform.
  • a mobile platform with a distance measuring device can measure the external environment, for example, measuring the distance between the mobile platform and obstacles for obstacle avoidance and other purposes, and for two-dimensional or three-dimensional mapping of the external environment.
  • the mobile platform includes at least one of an unmanned aerial vehicle, a car, a remote control car, a robot, a boat, and a camera.
  • the ranging device is applied to an unmanned aerial vehicle
  • the platform body is the fuselage of the unmanned aerial vehicle.
  • the distance measuring device is applied to a car
  • the platform body is the body of the car.
  • the car can be a self-driving car or a semi-autonomous car, and there is no restriction here.
  • the platform body is the body of the remote control car.
  • the platform body is a robot.
  • the distance measuring device is applied to a camera, the platform body is the camera itself.
  • the emission angle of the laser is constantly changing, but the emission angle of these lasers is not necessarily uniform in the scanning field of the lidar, and the difference can be several to hundreds of times, or even more.
  • the uneven scanning angle results in uneven density of point cloud images in different regions, that is, the distance measuring device has a non-uniform scanning density in the scanning field of view, that is, it has an irregular sampling pattern, thus causing the distance measuring device
  • the uniformity of the acquired initial point cloud is poor.
  • the scan density in the substantially central area is higher than other areas, and the point cloud image density of this area is also higher than other areas.
  • the uniformity and noise of the point cloud output by the laser radar ranging device are important parameters of the point cloud quality, which affect the display effect and further processing of the point cloud.
  • improving the quality of point clouds is usually done by improving hardware sampling and point cloud generation algorithms.
  • the above-mentioned solutions have higher requirements for the entire system, and usually ignore the other and are restricted by the difficulty of system construction, power consumption, cost, and scanning rate.
  • the point cloud output by the optimized design still has some quality problems.
  • the present invention provides a point cloud resampling method.
  • the method includes: resampling an initial point cloud acquired by a distance measuring device to obtain a resampled point cloud after resample.
  • the distance measuring device has a non-uniform scanning density in the scanning field of view, the distribution uniformity of the resampled point cloud is higher than that of the initial point cloud, and/or the noise of the resampled point cloud Noise below the initial point cloud.
  • the resampled point cloud can be more suitable for subsequent processing, and/or have better display effects, and/or hide specific hardware sampling modes, etc., and the noise of the resampled point cloud is lower than the initial point cloud The noise, thus improving the accuracy of ranging.
  • the point cloud resampling method of the present invention includes: re-sampling the initial point cloud acquired by the ranging device to obtain the resampled point cloud after re-sampling, wherein the ranging device is scanning
  • the field of view has a non-uniform scanning density, and the distribution uniformity of the resampled point cloud is higher than the distribution uniformity of the initial point cloud, as shown in Figure 4. Therefore, the resampled point cloud is more suitable for subsequent processing, for example Subsequent processing such as object recognition and image fusion has the advantages of better display and hiding specific hardware sampling modes.
  • the distribution uniformity of the resampled point cloud is higher than that of the initial point cloud after interpolation is performed at the point cloud gap of the initial point cloud.
  • the projection points of the resampled point cloud on the reference surface have a higher uniformity than the projection points of the initial point cloud on the reference surface.
  • the reference surface includes at least one of a plane, a spherical surface, a cylindrical surface, or other suitable reference surfaces.
  • the reference plane includes a plane, the plane being a plane perpendicular to the central axis of the light pulse sequence emitted by the distance measuring device, wherein the plane may also be the image plane of the distance measuring device , Or other suitable plane, such as a plane not perpendicular to the central axis of the light pulse sequence emitted by the distance measuring device.
  • the projection point of the initial point cloud or the resampled point cloud on the reference surface may refer to: taking the reference surface as the projection surface, and taking the position of the ranging device (such as lidar) as the center point, and The point cloud points of the initial point cloud or the resampled point cloud are respectively perspectively projected to the reference surface to obtain projection points respectively located on the reference surface.
  • the reference surface includes a spherical surface, for example, the spherical surface is a spherical surface centered on the distance measuring device.
  • the reference surface includes a cylindrical surface, for example, the central axis of the cylindrical surface is a vertical line passing through the distance measuring device, and the vertical line may be perpendicular to the central axis of the distance measuring device, for example. Or, the vertical line may be perpendicular to the horizontal plane, for example.
  • At least part of the projection points of the resampled point cloud on the reference surface are evenly distributed, for example, part of the projection points in the central area of the reference surface are evenly distributed; or, part of the projection points in other parts of the area Or, the projection points of the resampled point cloud on the entire reference surface are uniformly distributed; or, at least part of the projection points of the resampled point cloud on the reference surface are on the reference surface
  • At least one track is arranged at an equal angle, as shown in Figure 10, where the equal angle arrangement means that in at least part of the projection points, the angle formed by any two adjacent projection points and the line connecting the distance measuring device is the same
  • at least part of the projection points of the resampled point cloud on the reference surface are arranged at equal intervals on at least one trajectory of the reference surface. It is worth mentioning that the equally spaced rows For example, the linear distance between adjacent projection points is the same.
  • the trajectory includes a regular straight line, a regular curve, a regular broken line, or other suitable regular trajectories.
  • the track may also include a track extending in a horizontal direction and/or a vertical direction.
  • the track includes a track having a regular curve, and the track includes a track extending in a horizontal direction (generally a regular curve) and a track extending in a vertical direction (generally Is a regular straight line).
  • the track includes a regular linear track, and the track includes a track extending in a horizontal direction and a vertical direction.
  • the trajectory includes multiple concentric rings, such as concentric circular rings, concentric square rings, concentric triangular rings, concentric polygonal rings, etc., wherein, as shown in FIG. 14, the trajectory includes Multiple concentric rings.
  • the centers of the multiple concentric rings are located on the central axis of the distance measuring device, so that the trajectories are symmetrically distributed around the central axis, thereby ensuring uniform sampling along the trajectories.
  • the track includes a spiral track.
  • the trajectory includes multiple trajectories radiating outward along a central point.
  • interpolation is performed at the point cloud gap of the initial point cloud, so that the projection points of the resampled point cloud on the reference surface are more uniform than the projection points of the initial point cloud on the reference surface. Degree is high.
  • the target scene detected by the distance measuring device can be an outdoor scene or an indoor scene.
  • the same indoor scene is mainly used as an example in the embodiment of this application.
  • the indoor scene shown in FIG. There are square plates in the front, tables and seats on the left and right sides, chandeliers on the roof, and the ground.
  • Figure 8 shows the interpolated depth distribution of the indoor scene
  • Figures 3, 4, 7, 11 , 13, 15, and 17 are all point cloud images of the indoor scene.
  • Figures 3 and 4 are the initial point cloud images
  • Figure 7 is the discretized point cloud projection
  • Figure 15 and Figure 17 are resampled point clouds obtained by other different types of resampling methods.
  • any suitable interpolation method can be used to perform interpolation at the point cloud gap of the initial point cloud.
  • the initial point cloud may be reduced in dimensionality first, and the initial point cloud may be interpolated in low dimensions to reduce the difficulty of interpolation.
  • project a three-dimensional initial point cloud onto a two-dimensional reference surface to obtain an image on the reference surface, for example, obtain a point cloud projection map (or depth distribution map) on the reference surface.
  • a point cloud projection map or depth distribution map
  • the interpolation method includes the following steps A1 to A3: in step A1, the initial point cloud is projected onto a reference surface, Obtain the position of the projection point of the initial point cloud on the reference surface, and the value (or pixel value) of the reference point.
  • the value of the projection point includes the characteristic information of the corresponding initial point cloud point, such as depth information and/ Or reflectivity information.
  • the reference plane that is, the projection plane
  • the plane can also be the image plane of the distance measuring device, or other suitable planes.
  • the reference plane can be used as the projection plane, and the position of the distance measuring device (such as lidar) can be used as the center point.
  • the perspective projection of the cloud point to the reference surface obtains the projection point on the reference surface.
  • the projection diagram of the initial point cloud on the reference surface shown in Fig. 6 shows that the initial point cloud is sampled before resampling The density is not uniform, and the sampling density in the central area is significantly higher than other areas.
  • the interpolation method further includes step A2, discretizing the projection points of the initial point cloud on the reference surface, and this step can be selectively performed as required.
  • discretization includes using a pixel to replace one or more projection points of the initial point cloud that fall into the pixel.
  • the pixel value of the pixel is calculated according to the projection point of the initial point cloud falling into the pixel.
  • step A2 can also easily determine the position of the interpolation.
  • the discretization method includes: gridding the projection points of the initial point cloud on the reference surface, wherein the position of the interpolation point includes at least a part of an empty grid.
  • a pixel can include one or more grids.
  • the grid is formed by a plurality of mutually perpendicular trajectories on the reference surface.
  • the grid is formed by a plurality of traces extending in a horizontal direction and a plurality of traces extending in a vertical direction on the reference surface.
  • the grid may be a grid formed by a plurality of intersecting linear trajectories extending in a horizontal direction and a plurality of intersecting linear trajectories extending in a vertical direction.
  • the grid is formed by intersecting multiple concentric circles on the reference plane and multiple tracks radiating outward from the center of the concentric circles.
  • the grid is formed by a spiral line on the reference surface intersecting multiple tracks radiating from the center of the spiral line. The size between adjacent trajectories can be set reasonably according to needs, thereby setting the physical size of the grid.
  • the method of gridding includes: substituting the depth value of the grid for the depth values of a plurality of projection points in the grid, wherein the depth value of the grid is based on falling into the grid The depth values of multiple projection points are determined.
  • each grid can correspond to a pixel.
  • the depth value of the grid is also the pixel value of the pixel. Due to the physical size of the grid, there may be multiple projection points on the same grid. In the case of medium, the depth value of the grid can be determined based on the depth values of multiple projection points that fall into the grid, for example, can be based on the minimum depth of multiple projection points that fall into the grid The value is determined, or it can be based on the average depth value of multiple projection points that fall into the grid. In a grid that only includes a unique projection point, the depth value of the grid is also the depth value of the projection point.
  • the interpolation method further includes step A3, determining interpolation points at the gaps of the projection points on the reference surface.
  • the depth value of the interpolation point may be determined based on any suitable interpolation method, for example, the depth value of the interpolation point is determined based on the depth value of the adjacent projection point, and further, the depth value of the interpolation point is based on The depth value of the nearest neighboring projection point is determined, or is determined based on the minimum depth value of the neighboring projection point.
  • the method for determining the interpolation point further includes: determining whether the position of the empty grid is a sky position, wherein the empty grid to be interpolated includes an empty grid of a non-sky position.
  • different values are assigned to the empty grid of the sky position and the empty grid of the non-sky position, for example, special values are assigned to the empty grid of the sky position.
  • the special value is a value different from the pixel value of the grid with projection points, for example, the special value is 0, or other suitable values.
  • the sky position in the reference plane can be judged according to the characteristics of the distance measuring device when detecting the sky position and the characteristics of the sky position itself. There are two main types: 1) The sky position is relatively large, that is, the area is larger; 2) When the distance device sends out a detection signal for detection, there is a transmission signal but no reception signal in the sky position.
  • the method for interpolating the empty grids of non-sky positions includes: interpolating the empty grids of non-sky positions based on a neighborhood interpolation algorithm. All conventional field interpolation algorithms in digital images can be applied to the present invention , Neighborhood interpolation algorithms include nearest neighbor interpolation algorithm or linear interpolation algorithm to assign suitable depth values to the empty grids of non-sky positions, as shown in Figure 8 is the depth distribution map obtained after interpolation, thus obtaining a sample Uniform image.
  • the interpolated point cloud can be resampled in any mode on the basis of the image to obtain the resampled point cloud with a specific arrangement pattern, thereby providing more point cloud presentation results , This step can be carried out selectively as required.
  • point cloud resampling can be performed on a two-dimensional reference surface. In an example, the position of the projection points on the two-dimensional reference surface is modified so that the resampled projection points are arranged in a specific pattern.
  • the pixel value of the modified projection point (that is, the resampled projection point) can be determined according to the pixel value of the projection point at the same position on the reference surface before resampling, or according to the same pixel value on the reference surface before resampling.
  • the pixel value of the adjacent projection point at the position is determined, for example, based on interpolation of the pixel value of the adjacent projection point at the same position on the reference surface before resampling, where the pixel value in this document includes depth information and/or reflectance information.
  • the scanning direction ( ⁇ , ) are uniformly sampled respectively, where ⁇ and Corresponds to the coordinates in the polar coordinate system as shown in Figure 9, uniform sampling such as uniformly changing ⁇ and Then, for each re-sampling direction, the intersection of the re-sampling direction and the image on the reference surface (such as the image shown in Fig. 8) is calculated, and the value of the intersection is obtained by, for example, the aforementioned interpolation method, so that the re-sampling direction can be obtained. The pixel value in the sampling direction is then determined to determine a resampled point cloud point.
  • the nearest neighbor interpolation method can be used, and the value of the pixel (that is, the grid) where the intersection point is located is used as the pixel value of the sampling direction.
  • the resampled point cloud as shown in FIG. 11 can be obtained, and at least part of the projection points of the resampled point cloud on the reference surface are equiangular on at least one track of the reference surface.
  • the projection point corresponds to the intersection of the re-sampling direction set and the image on the reference surface.
  • the point cloud resampling method can also be called a planar uniform resampling method. For example, traverse the aforementioned interpolated image (that is, For each pixel of the image on the reference surface, if the pixel value of the pixel is non-zero, a re-sampling point is generated.
  • the re-sampling point can be located at the origin (for example, the origin of the distance measuring device is the origin) and the center of the image pixel
  • the depth value is the pixel value of the corresponding pixel to obtain the resampled point cloud. As shown in FIG.
  • the distribution uniformity of the resampled point cloud is higher than that of the initial point cloud.
  • At least some of the projection points of the resampled point cloud on the reference surface are arranged at equal intervals on at least one track of the reference surface.
  • Figure 12 shows the projection point of the resampled point cloud on the reference surface (the projection point corresponds to the center of the image pixel).
  • the projection point of the resampled point cloud on the reference surface is compared to the original point cloud.
  • the uniformity of the projection points on the reference surface is high, so that a plane uniformly resampled point cloud can be obtained.
  • the method of the foregoing embodiment to first generate uniform points on the plane to be sampled, which are arranged at equal intervals, and then calculate the intersection points between the uniform points and the image on the reference surface, and then use the interpolation method. Way to obtain the resampled point cloud.
  • re-sampling according to a specific pattern first generates a set of re-sampling directions according to the requirements of the target pattern, and these sets of re-sampling directions are in the reference
  • the planes are arranged along a trajectory, for example, as shown in FIG. 14, the trajectory includes a plurality of concentric circular rings (that is, uniform circumferential sampling), or, as shown in FIG. 16, the trajectory includes a spiral trajectory.
  • the distribution uniformity of the resampled point cloud obtained by the above resampling method is higher than the distribution uniformity of the initial point cloud, and the initial point cloud with irregular sampling mode can be converted into a uniformly sampled resampled point cloud.
  • the point cloud can be more suitable for subsequent processing (such as object recognition, image fusion, etc.), and/or have better display effects, and/or hide specific hardware sampling modes.
  • the resampled point cloud adds a plurality of point cloud points located on the same reference surface, that is, the resampled point cloud increases compared to the initial point cloud Multiple projection points located on the same reference plane.
  • the method may further include: adding a determined interpolation point to the initial point cloud.
  • a new re-sampling direction is added to the initial point cloud sampling direction to increase the sampling density, where the point cloud point corresponding to each re-sampling direction is used as the re-sampling point cloud point, which can be achieved by, for example, interpolation Calculate the pixel values of the resampled point cloud points.
  • the interpolation method may be based on the interpolation method in the foregoing embodiment, such as the nearest neighbor interpolation algorithm or the linear interpolation algorithm.
  • the nearest neighbor interpolation algorithm or the linear interpolation algorithm.
  • different areas of the initial point cloud have different point cloud densities
  • re-sampling the initial point cloud includes: down-sampling the initial point cloud in an area with high point cloud density .
  • the scanning density of the distance measuring device in the central area of the scanning field of view is higher than the scanning density of other areas
  • the re-sampling of the initial point cloud includes: down-sampling the initial point cloud in the central area to Equalize the sampling density.
  • the method of downsampling includes randomly or uniformly discarding all initial point cloud points corresponding to a part of the grid in the central area, wherein the grid is obtained by dividing the initial point on a reference surface The projection points of the cloud are gridded.
  • the method of downsampling includes: limiting the number of initial point cloud points that fall into at least a part of the grid in the central area to be lower than a threshold number, wherein the grid is generated by The projection points of the initial point cloud are gridded on the reference surface.
  • the method for restricting the number of initial point cloud points in at least part of the grid falling in the central area to be less than a threshold number includes: randomly or uniformly discarding part of the initial point cloud in the at least part of the grid Points (that is, the projection points of the part of the initial point cloud point in the part of the grid are discarded) until the number of projection points in the at least part of the grid is within the threshold.
  • the threshold number is set reasonably according to actual needs, for example, the number of projection points in all grids in the central area is divided by the number of grids in the central area to obtain an average value, and the threshold number is set equal to the average Value, the number of projection points in at least part of the grid in the control center area is lower than the threshold, or the number of projection points of all initial point clouds on the reference surface is divided by the total number of grids to obtain an average value, The threshold number is set equal to the average value.
  • the above two threshold setting methods are only examples. Other thresholds that can equalize the number of point clouds in areas with high sampling density can also be applied to the present invention.
  • re-sampling the initial point cloud includes: denoising and re-sampling the initial point cloud, so that the noise of the re-sampled point cloud is lower than that of the initial point cloud Noise.
  • the method for noise reduction and resampling includes: performing noise filtering processing on an image of the initial point cloud on a reference surface.
  • the processing should be edge-preserving.
  • a bilateral filter is used for noise filtering, and the image after noise filtering has good continuity and easy observation.
  • the image may include a depth distribution map or a reflectance distribution map.
  • the image is obtained based on the method in the foregoing embodiment.
  • the image is mainly used as the depth distribution map.
  • the method of resampling is described.
  • the method of obtaining the image includes: projecting the initial point cloud onto a reference surface to obtain the projection points of the initial point cloud, and these projections
  • the point includes the characteristic information of the corresponding initial point cloud point, such as depth information and/or reflectivity information, as shown in FIG.
  • the projection points of the initial point cloud are gridded on the reference surface ,
  • each grid corresponds to a pixel
  • the gridization process refers to the description in the foregoing embodiment, which will not be repeated here; interpolation is performed on part of the empty grid.
  • the pixel value of the grid is determined according to the pixel values of the multiple projection points that fall into the grid, for example, the pixel value of the grid is the smallest pixel value of the projection point that falls into the grid; The pixel value of the grid determines the pixel value of the empty grid.
  • the method for determining the pixel value of the empty grid based on the pixel value of the adjacent grid is specifically: determining whether the position of the empty grid is a sky position, wherein the determination of the non-empty grid is based on the pixel value of the adjacent grid.
  • the pixel value of the empty grid of the sky position is specifically: determining whether the position of the empty grid is a sky position, wherein the determination of the non-empty grid is based on the pixel value of the adjacent grid.
  • the method for noise reduction and resampling further includes: traversing each initial point cloud point of the initial point cloud; and mapping each initial point cloud point to the noise-filtered Image; adjust the pixel value of each initial point cloud point to the pixel value of its corresponding pixel, or adjust to the pixel value of the adjacent pixel of the initial point cloud point in the projection position of the image, in the corresponding
  • the pixel value of the neighboring pixel may also be the pixel value after interpolation.
  • the method for noise reduction and resampling further includes: performing threshold processing based on the adjusted pixel value of the initial point cloud point and the initial pixel value of the initial point cloud point, and the initial pixel value and When the difference between the adjusted pixel values is greater than the target threshold, it indicates that the initial point cloud point is likely to be a noise point, so the initial point cloud point is discarded.
  • noise on the edge there may be multiple point cloud points in the same sampling direction in the initial point cloud (the larger depth is basically It can be considered as noise) and points that roughen the surface of the plane, which in turn makes the surface of the plane more flat and the edges of the object clearer, which facilitates the application of subsequent processing and improves the measurement accuracy of the distance measuring device.
  • a point cloud resampling method is also provided.
  • the point cloud resampling method mainly includes the following steps:
  • the initial point cloud is projected onto a reference surface to obtain the projection points of the initial point cloud.
  • the description of the initial point cloud refers to the foregoing embodiment.
  • the reference surface includes at least one of a plane, a spherical surface, and a cylindrical surface.
  • the reference surface includes a two-dimensional plane as an example for explanation and description; then, an interpolation point is determined at the gap between the projection points on the reference surface to obtain
  • the image above which may be a two-dimensional image on a two-dimensional plane, such as a two-dimensional depth distribution map, etc., and then resample based on the image to obtain a resampled point cloud with a specific arrangement pattern, or Perform noise reduction and resampling based on the image, so that the noise of the resampled point cloud is lower than the noise of the initial point cloud.
  • the image above which may be a two-dimensional image on a two-dimensional plane, such as a two-dimensional depth distribution map, etc., and then resample based on the image to obtain a resampled point cloud with a specific arrangement pattern, or Perform noise reduction and resampling based on the image, so that the noise of the resampled point cloud is lower than the noise of the initial point cloud.
  • the projection point of the initial point cloud can be discretized on the reference surface.
  • the specific steps of the discretization refer to the foregoing embodiment, and will not be repeated here. .
  • the resampled point cloud of a specific arrangement pattern can refer to the above description of the resampled point cloud, for example, any one of the rows in Figure 11, Figure 13, Figure 15 and Figure 17 in the foregoing embodiment. Resample point cloud of cloth pattern.
  • the distribution uniformity of the resampled point cloud obtained based on the method of this embodiment is higher than the distribution uniformity of the initial point cloud, and/or the noise of the resampled point cloud is lower than that of the initial point cloud noise.
  • the initial point cloud with irregular sampling patterns can be converted into a uniformly sampled resampled point cloud, and the sampling density of the point cloud can be increased or decreased.
  • the resampled point cloud can be more suitable for subsequent processing, and/or have better display effects, and/or hide specific hardware sampling modes, etc., and the noise of the resampled point cloud is lower than the initial point cloud The noise, thus improving the accuracy of ranging.
  • the embodiment of the present invention also provides a point cloud resampling device 500.
  • the sampling device 500 includes a re-sampling module 510 for re-sampling the initial point cloud acquired by the ranging device to obtain a re-sampled point cloud after re-sampling, wherein the ranging device has a non-uniform scanning field of view.
  • the scanning density of the resampled point cloud is higher than that of the initial point cloud, and/or the noise of the resampled point cloud is lower than the noise of the initial point cloud.
  • the projection points of the resampled point cloud on the reference surface have a higher uniformity than the projection points of the initial point cloud on the reference surface.
  • the reference surface includes at least one of a flat surface, a spherical surface, and a cylindrical surface.
  • the plane is a surface perpendicular to the central axis of the light pulse sequence emitted by the distance measuring device; or, the spherical surface is a spherical surface centered on the distance measuring device; or, the cylindrical surface
  • the central axis is the vertical line passing through the distance measuring device.
  • the projection points of the resampled point cloud on the reference surface are arranged at equal intervals or at equal angles on at least one track of the reference surface.
  • the trajectory includes a trajectory in a regular straight line, a regular curve, or a regular broken line, or other suitable trajectories.
  • the track includes a track extending in a horizontal direction and/or a vertical direction.
  • the trajectory includes a plurality of concentric rings, and the plurality of concentric rings may have the same interval or different intervals.
  • the centers of the multiple concentric rings are located on the central axis of the distance measuring device.
  • the track includes a spiral track.
  • the trajectory includes multiple trajectories radiating outward along a central point.
  • the resampled point cloud compared with the initial point cloud, adds a plurality of point cloud points located on the same reference surface. Therefore, the uniformity of the resampled point cloud is higher.
  • the point cloud resampling device further includes an image generation module, the image generation module includes a projection module and an interpolation module, the projection module is used to project the initial point cloud onto a reference surface to obtain the projection points of the initial point cloud
  • the interpolation module is used to perform interpolation at the point cloud gap of the initial point cloud, more specifically, the interpolation module is used to determine the interpolation point at the gap of the projection point on the reference surface to obtain
  • the value of the interpolation point is determined based on the pixel value (for example, the depth value) of the adjacent projection point, or other suitable interpolation methods can also be applied to the present invention .
  • the image generation module further includes a discretization module, configured to discretize the projection points of the initial point cloud on the reference surface before determining the interpolation point.
  • the discretization module is specifically configured to: grid the projection points of the initial point cloud on the reference surface, wherein the position of the interpolation point includes at least part of an empty grid. Substituting the pixel values of the grid for the pixel values of the multiple projection points in the grid, where the pixel values of the grid are determined according to the pixel values of the multiple projection points that fall into the grid, then The value of the interpolation point may be determined based on the pixel value of the adjacent grid.
  • the grid is formed by a plurality of mutually perpendicular trajectories on the reference surface, for example, the grid is formed by a plurality of traces extending in a horizontal direction and a plurality of traces extending in a vertical direction on the reference surface. Formed by crossing tracks.
  • the grid may also be formed by intersecting multiple concentric circular rings on the reference surface and multiple trajectories radiating outward from the center of the concentric ring.
  • the grid may also be formed by a spiral line on the reference surface intersecting multiple tracks radiating from the center of the spiral line.
  • the interpolation module is also used to determine whether the position of the empty grid is a sky position, wherein the empty grid to be interpolated includes an empty grid of a non-sky position.
  • different values are assigned to the empty grid of the sky position and the empty grid of the non-sky position, for example, special values are assigned to the empty grid of the sky position.
  • the special value is a value different from the pixel value of the grid with projection points, for example, the special value is 0, or other suitable values.
  • the sky position in the reference plane can be judged according to the characteristics of the distance measuring device when detecting the sky position and the characteristics of the sky position itself. There are two main types: 1) The sky position is relatively large, that is, the area is larger; 2) When the distance device sends out a detection signal for detection, there is a transmission signal but no reception signal in the sky position.
  • the method for interpolating the empty grids of non-sky positions includes: interpolating the empty grids of non-sky positions based on a neighborhood interpolation algorithm.
  • the neighborhood interpolation algorithm includes the nearest neighbor interpolation algorithm or linear interpolation algorithm, etc., to assign suitable pixel values to the empty grid of non-sky positions, thereby obtaining a uniformly sampled image, and the resampling module is also specifically used for: Resampling is performed based on the image to obtain the resampled point cloud with a specific arrangement pattern.
  • the re-sampling module is specifically configured to perform re-sampling based on the image to obtain the re-sampling point cloud with a specific arrangement pattern.
  • the process of obtaining the re-sampling point cloud with a specific arrangement pattern reference may be made to the description in the foregoing embodiment, which is not repeated here.
  • the resampling module is further configured to: add a certain interpolation point to the initial point cloud. Compared with the initial point cloud, the resampled point cloud adds multiple point cloud points located on the same reference surface, that is, compared to the initial point cloud, the resampled point cloud adds multiple points located on the same reference surface. The projection point on the reference surface.
  • a new re-sampling direction is added to the initial point cloud sampling direction to increase the sampling density, where the point cloud point corresponding to each re-sampling direction is used as the re-sampling point cloud point, which can be achieved by, for example, interpolation Calculate the pixel values (such as depth values) of the resampled point cloud points.
  • the interpolation method may be based on the interpolation method in the foregoing embodiment, such as the nearest neighbor interpolation algorithm or the linear interpolation algorithm.
  • the nearest neighbor interpolation algorithm or the linear interpolation algorithm.
  • the resampling module further includes a downsampling module for downsampling the initial point cloud in areas with high point cloud density.
  • the scanning density of the distance measuring device in the central area of the scanning field of view is higher than the scanning density of other areas, and the down-sampling module is specifically configured to down-sample the initial point cloud in the central area.
  • the downsampling module is specifically configured to randomly or uniformly discard all initial point cloud points corresponding to the partial grid of the central area, wherein the grid is obtained by dividing the The projection points of the initial point cloud are gridded.
  • the down-sampling module is specifically configured to limit the number of initial point cloud points falling in at least a part of the grid in the central area to be lower than a threshold number, wherein the grid is generated by The projection points of the initial point cloud are gridded on the reference surface.
  • the down-sampling module is specifically used to randomly or uniformly discard part of the initial point cloud points in the at least part of the grid.
  • the re-sampling module further includes a noise reduction and re-sampling module, configured to perform noise reduction and re-sampling on the initial point cloud, so that the noise of the re-sampled point cloud is lower than the initial point The noise of clouds.
  • the noise reduction and resampling module is specifically configured to perform noise filtering processing on the image of the initial point cloud on the reference surface.
  • the noise reduction and resampling module is specifically further used to: traverse each initial point cloud point of the initial point cloud; map each initial point cloud point to all the points after the noise filtering process.
  • the image adjust the pixel value (for example, the depth value) of each initial point cloud point to the pixel value of its corresponding pixel, or adjust to the adjacent pixel of the initial point cloud point at the projection position of the image The pixel value.
  • the noise reduction and resampling module is further configured to: perform threshold processing based on the adjusted pixel value of the initial point cloud point and the initial pixel value of the initial point cloud point, and after the initial pixel value and the adjusted When the pixel value difference of is greater than the target threshold, the initial point cloud point is discarded.
  • the image includes a depth distribution map, or other suitable images, such as a reflectance distribution map.
  • the point cloud resampling device further includes an image generation module for obtaining the image, the image generation module includes a discretization module and an interpolation module, and the discretization module is used to compare all the images on the reference surface.
  • the projection points of the initial point cloud are gridded, where each grid corresponds to a pixel, and the discretization module is also used to determine the grid’s size according to the pixel values of multiple projection points that fall into the grid. Pixel value; and; the interpolation module is used to interpolate a part of the empty grid, more specifically, the interpolation module is used to determine the pixel value of the empty grid based on the pixel value of the adjacent grid.
  • the interpolation module is specifically further configured to determine whether the position of the empty grid is a sky position, wherein the pixel value of the empty grid of a non-sky position is determined based on the pixel value of an adjacent grid.
  • the point cloud resampling method according to the embodiment of the present invention may be implemented in a device, apparatus, or system including a memory and a processor.
  • FIG. 19 shows a schematic block diagram of a point cloud resampling system 600 in an embodiment of the present invention.
  • the point cloud resampling system 600 includes one or more ranging devices 610, one or more processors 630, and one or more storage devices 620.
  • the point cloud resampling system 600 may also include at least one of an input device (not shown), an output device (not shown), and an image sensor (not shown). These components are connected through a bus system and/or other components. A form of connection mechanism (not shown) is interconnected.
  • the components and structure of the point cloud resampling system 600 shown in FIG. 19 are only exemplary and not restrictive. According to needs, the point cloud resampling system 600 may also have other components and structures, for example, It may include a transceiver for transceiving signals.
  • the storage device 620 is also a memory for storing processor-executable instructions, for example, for storing corresponding steps and program instructions in the point cloud resampling method according to the embodiment of the present invention. It may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and/or non-volatile memory.
  • the volatile memory may include random access memory (RAM) and/or cache memory (cache), for example.
  • the non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.
  • the input device may be a device used by a user to input instructions, and may include one or more of a keyboard, a mouse, a microphone, and a touch screen.
  • the output device may output various information (for example, images or sounds) to the outside (for example, a user), and may include one or more of a display, a speaker, and the like.
  • the communication interface (not shown) is used for communication between the point cloud resampling system 600 and other devices, including wired or wireless communication.
  • the point cloud resampling system 600 can access a wireless network based on a communication standard, such as WiFi, 2G, 3G, 4G, 5G, or a combination thereof.
  • the communication interface receives a broadcast signal or broadcast related information from an external broadcast management system via a broadcast channel.
  • the communication interface further includes a near field communication (NFC) module to facilitate short-range communication.
  • the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.
  • RFID radio frequency identification
  • IrDA infrared data association
  • UWB ultra-wideband
  • Bluetooth Bluetooth
  • the processor 630 may be a central processing unit (CPU), an image processing unit (GPU), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other forms with data processing capabilities and/or instruction execution capabilities
  • the processing unit of the point cloud and can control other components in the point cloud resampling system 600 to perform desired functions.
  • the processor can execute the instructions stored in the storage device 620 to execute the point cloud resampling method described herein.
  • the processor 630 can include one or more embedded processors, processor cores, microprocessors, logic circuits, hardware finite state machines (FSM), digital signal processors (DSP), or combinations thereof.
  • One or more computer program instructions can be stored on the computer-readable storage medium, and the processor 630 can run the program instructions stored in the storage device 620 to implement the embodiments of the present invention described herein (implemented by the processor). ) And/or other desired functions, for example, to perform the corresponding steps of the point cloud resampling method according to the embodiment of the present invention, and to implement each module in the point cloud resampling device according to the embodiment of the present invention .
  • Various application programs and various data such as various data used and/or generated by the application program, can also be stored in the computer-readable storage medium.
  • the embodiment of the present invention also provides a computer storage medium on which a computer program is stored.
  • the computer program is executed by the processor, each step of the point cloud data-based point cloud resampling method of the embodiment of the present invention or each component module in the aforementioned point cloud resampling device can be realized.
  • the computer storage medium may include, for example, a memory card of a smart phone, a storage component of a tablet computer, a hard disk of a personal computer, a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a portable compact disk Read only memory (CD-ROM), USB memory, or any combination of the above storage media.
  • the computer-readable storage medium may be any combination of one or more computer-readable storage media.
  • one computer-readable storage medium contains computer-readable program codes for converting the point cloud data into a two-dimensional image
  • another computer-readable storage medium contains a computer for object segmentation of the two-dimensional image. Readable program code, etc.
  • the disclosed device and method may be implemented in other ways.
  • the device embodiments described above are merely illustrative.
  • the division of the units is only a logical function division, and there may be other divisions in actual implementation, for example, multiple units or components can be combined or It can be integrated into another device, or some features can be ignored or not implemented.
  • the various component embodiments of the present invention may be implemented by hardware, or by software modules running on one or more processors, or by a combination of them.
  • a microprocessor or a digital signal processor (DSP) may be used in practice to implement some or all of the functions of some modules according to the embodiments of the present invention.
  • DSP digital signal processor
  • the present invention can also be implemented as a device program (for example, a computer program and a computer program product) for executing part or all of the methods described herein.
  • Such a program for realizing the present invention may be stored on a computer-readable medium, or may have the form of one or more signals. Such signals can be downloaded from Internet websites, or provided on carrier signals, or provided in any other form.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Optical Radar Systems And Details Thereof (AREA)

Abstract

一种点云重采样的方法、装置（500）和系统（600），点云重采样的方法包括：对测距装置（100，200，610）获取的初始点云进行重采样，以获得重采样后的重采样点云，测距装置（100，200，610）在扫描视场中具有非均匀的扫描密度，重采样点云的分布均匀性高于初始点云的分布均匀性，和/或，重采样点云的噪声低于初始点云的噪声。通过点云重采样的方法，可以将具有非规则采样模式的初始点云转换为均匀采样的重采样点云，可以增加或减少点云的采样密度，重采样后的点云，可以更适合后续处理，和/或拥有更好的显示效果，和/或隐藏特定的硬件采样模式等，以及重采样点云的噪声低于初始点云的噪声，因此提高了测距准确性。

Description

一种点云重采样的方法、装置和系统
说明书
技术领域
本发明总地涉及测距装置技术领域，更具体地涉及一种点云重采样的方法、装置和系统。
背景技术
激光雷达是一种外界的感知系统，通过发射并接收处理光波信息可以获知外界的立体三维信息，不再局限于相机等对外界的平面感知方式。例如激光雷达的测距装置输出的点云的均匀度、噪点等是点云质量的重要参数，影响着点云的显示效果、进一步处理等。目前改善点云质量，通常通过改善硬件采样以及点云生成算法来进行。然而上述方案对整个系统的要求较高，通常会顾此失彼，受到系统搭建难度、功耗、成本、扫描速率等的制约。在实际中，经优化设计的系统所输出的点云仍总是存在一些质量问题。
发明内容
为了解决上述问题中的至少一个而提出了本发明。具体地，本发明一方面提供一种点云重采样的方法，所述方法包括：
对测距装置获取的初始点云进行重采样，以获得重采样后的重采样点云，其中，所述测距装置在扫描视场中具有非均匀的扫描密度，所述重采样点云的分布均匀性高于所述初始点云的分布均匀性，和/或，所述重采样点云的噪声低于所述初始点云的噪声。
示例性地，所述对测距装置输出的初始点云进行重采样包括：
在所述初始点云的点云间隙处进行插值。
示例性地，所述重采样点云在参考面上的投影点相比所述初始点云在该参考面上的投影点的均匀度高。
示例性地，所述参考面包括平面、球面、柱面中的至少一种。
示例性地，所述平面为与所述测距装置所发射的光脉冲序列的中心轴相垂直的面；或者，
所述球面为以所述测距装置为中心的球面；或者，
所述柱面的中轴线为经过所述测距装置的竖直线。
示例性地，所述重采样点云在所述参考面上的至少部分投影点均匀分布。
示例性地，所述重采样点云在所述参考面上的投影点中的至少部分投影点在该参考面的至少一条轨迹上等间隔排布或者等角度排布。
示例性地，所述轨迹包括呈规则的直线状、规则的曲线状或者规则的折线状的轨迹。
示例性地，所述轨迹包括沿水平方向和/或竖直方向上延伸的轨迹。
示例性地，所述轨迹包括多个同心环。
示例性地，所述多个同心环的中心位于所述测距装置的中心轴上。
示例性地，所述轨迹包括一条呈螺旋线的轨迹。
示例性地，所述轨迹包括多条沿一个中心点往外辐射的多条轨迹。
示例性地，所述重采样点云相比所述初始点云增加了多个位于同一个所述参考面上的点云点。
示例性地，所述插值的方法包括：将所述初始点云投影到参考面上，得到所述初始点云的投影点；
在所述参考面上的所述投影点的间隙处确定插值点，以获得位于所述参考面上的图像。
示例性地，所述插值点的深度值是基于邻近的所述投影点的深度值确定的。
示例性地，所述插值的方法还包括：在确定所述插值点之前，在所述参考面上将所述初始点云的投影点进行离散化。
示例性地，所述离散化的方法包括：
在所述参考面上将所述初始点云的投影点进行网格化，其中，所述插值点的位置包括至少部分空网格。
示例性地，所述对测距装置获取的初始点云进行重采样包括：
基于所述图像进行重采样，以获得具有特定排布模式的所述重采样点云。
示例性地，所述网格化的方法包括：
将所述网格的深度值取代该网格内的多个投影点的深度值，其中该网格的深度值是根据落入到该网格中的多个投影点的深度值确定的；以及
所述插值点的深度值是基于邻近网格的深度值确定的。
示例性地，确定所述插值点的方法还包括：
确定所述空网格的位置是否天空位置，其中，待插值的所述空网格包括非天空位置的空网格。
示例性地，所述网格是由所述参考面上多条相互垂直的轨迹形成的。
示例性地，所述网格是由参考面上多条沿水平方向延伸和多条沿竖直方向延伸的交叉轨迹形成的。
示例性地，所述网格是由参考面上的多条同心圆环与从该同心环的中心向外辐射的多条轨迹交叉形成的。
示例性地，所述网格是由参考面上的一条螺旋线，与从该螺旋线的中心向外辐射的多条轨迹交叉形成的。
示例性地，所述方法还包括：
在所述初始点云中增加确定的插值点。
示例性地，所述初始点云不同区域的点云密度不同，对所述初始点云进行重采样包括：
在点云密度高的区域对所述初始点云进行降采样。
示例性地，所述测距装置在扫描视场的中心区域的扫描密度高于其余区域的扫描密度，对所述初始点云进行重采样包括：对中心区域的初始点云进行降采样。
示例性地，所述降采样的方法包括：
随机或者均匀舍弃所述中心区域的部分网格对应的全部初始点云点，其中，所述网格是由在参考面上将所述初始点云的投影点进行网格化而形成的。
示例性地，所述降采样的方法包括：
限制落入所述中心区域内的至少部分网格内的初始点云点的数量低于阈值数量，其中，所述网格是由在参考面上将初始点云的投影点进行网格化而形成的。
示例性地，所述限制落入所述中心区域内的至少部分网格内的初始点云点的数量低于阈值数量的方法，包括：
随机或者均匀舍弃所述至少部分网格内的部分初始点云点。
示例性地，对所述初始点云进行重采样包括：对所述初始点云进行降噪重采样，以使所述重采样点云的噪声低于所述初始点云的噪声。
示例性地，所述降噪重采样的方法包括：
对所述初始点云在参考面上的图像进行滤噪处理。
示例性地，所述降噪重采样的方法还包括：
遍历所述初始点云的每个初始点云点；
将每个所述初始点云点映射至所述滤噪处理后的所述图像；
将每个所述初始点云点的深度值调整为其对应像素的深度值，或者，调整为所述初始点云点在所述图像的投影位置的相邻像素的深度值。
示例性地，所述降噪重采样的方法还包括：
基于所述初始点云点调整后的深度值与该初始点云点的初始深度值进行阈值处理，在所述初始深度值和所述调整后的深度值之差大于目标阈值时，舍弃该初始点云点。
示例性地，所述图像包括深度分布图。
示例性地，获得所述图像的方法包括：
在所述参考面上将所述初始点云的投影点进行网格化，其中，每个网格对应一个像素；
对部分空网格进行插值。
示例性地，获得所述图像的方法包括：
根据落入到网格中的多个投影点的深度值确定该网格的深度值；以及
基于邻近网格的深度值确定空网格的深度值。
示例性地，所述基于邻近网格的深度值确定空网格的深度值的方法，具体为：
确定所述空网格的位置是否天空位置，其中，基于邻近网格的深度值确定非天空位置的所述空网格的深度值。
示例性地，所述测距装置包括用于出射光脉冲序列的发射模块，以及用于改变光脉冲序列的方向以对视场进行扫描的扫描模块。
示例性地，所述扫描模块在至少部分不同时刻的扫描路径不同。
示例性地，所述扫描模块在测距装置的视场内的扫描区域随着时间的累积而增加。
示例性地，所述扫描模块包括至少一个具有非平行的出射面和入射面的光折射元件。
示例性地，所述扫描模块包括在所述光脉冲序列的出射光路上依次排布的2个或3个所述光折射元件。
示例性地，所述扫描模块中的至少2个所述光折射元件在扫描过程中旋转，以改变所述光脉冲序列的方向。
本发明再一方面提供一种点云重采样装置，所述点云重采样装置包括：重采样模块，用于对测距装置获取的初始点云进行重采样，以获得重采样后的重采样点云，其中，所述测距装置在扫描视场中具有非均匀的扫描密度，所述重采样点云的分布均匀性高于所述初始点云的分布均匀性，和/或，所述重采样点云的噪声低于所述初始点云的噪声。
本发明另一方面提供一种点云重采样系统，点云重采样系统包括：
至少一个测距装置，用于探测目标场景生成初始点云；
存储器，用于存储可执行指令；
处理器，用于执行所述存储器中存储的所述指令，使得所述处理器执行前述的点云重采样的方法。
通过上述方法，可以将具有非规则采样模式的初始点云转换为均匀采样的重采样点云，可以增加或减少点云的采样密度。重采样后的点云，可以更适合后续处理，和/或拥有更好的显示效果，和/或隐藏特定的硬件采样模式等，以及所述重采样点云的噪声低于所述初始点云的噪声，因此提高了测距准确性。
附图说明
为了更清楚地说明本发明实施例中的技术方案，下面将对实施例描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。
图1示出了本发明一实施例中的测距装置的架构示意图；
图2示出了本发明一个实施例中的测距装置的示意图；
图3示出了本发明一个实施例中的测距装置所获得的初始点云的示意图；
图4示出了本发明一个实施例中的初始点云和重采样点云的对比示意图；
图5示出了本发明一个实施例中的点云投影示意图；
图6示出了本发明一个实施例中的初始点云在参考面上的投影示意图；
图7示出了本发明一个实施例中的离散化后的点云投影图；
图8示出了本发明一个实施例中的插值后的深度分布图；
图9示出了本发明一个实施例中的采样坐标系的示意图；
图10示出了本发明一个实施例中的重采样点云在参考面上的投影点的示意图；
图11示出了本发明一个实施例中的重采样点云的示意图；
图12示出了本发明另一个实施例中的重采样点云在参考面上的投影点的示意图；
图13示出了本发明另一个实施例中的重采样点云的示意图；
图14示出了本发明再一个实施例中的重采样点云在参考面上的投影点的示意图；
图15示出了本发明再一个实施例中的重采样点云的示意图；
图16示出了本发明又一个实施例中的重采样点云在参考面上的投影点的示意图；
图17示出了本发明又一个实施例中的重采样点云的示意图；
图18示出了本发明一个实施例中的点云重采样装置的示意性框图；
图19示出了本发明一个实施例中的点云重采样系统的示意性框图。
具体实施方式
为了使得本发明的目的、技术方案和优点更为明显，下面将参照附图详细描述根据本发明的示例实施例。显然，所描述的实施例仅仅是本发明的一部分实施例，而不是本发明的全部实施例，应理解，本发明不受这里描述的示例实施例的限制。基于本发明中描述的本发明实施例，本领域技术人员在没有付出创造性劳动的情况下所得到的所有其它实施例都应落入本发明的保护范围之内。
在下文的描述中，给出了大量具体的细节以便提供对本发明更为彻底的理解。然而，对于本领域技术人员而言显而易见的是，本发明可以无需一个或多个这些细节而得以实施。在其他的例子中，为了避免与本发明发生混淆，对于本领域公知的一些技术特征未进行描述。
应当理解的是，本发明能够以不同形式实施，而不应当解释为局限于这里提出的实施例。相反地，提供这些实施例将使公开彻底和完全，并且将本发明的范围完全地传递给本领域技术人员。
在此使用的术语的目的仅在于描述具体实施例并且不作为本发明的限制。 在此使用时，单数形式的“一”、“一个”和“所述/该”也意图包括复数形式，除非上下文清楚指出另外的方式。还应明白术语“组成”和/或“包括”，当在该说明书中使用时，确定所述特征、整数、步骤、操作、元件和/或部件的存在，但不排除一个或更多其它的特征、整数、步骤、操作、元件、部件和/或组的存在或添加。在此使用时，术语“和/或”包括相关所列项目的任何及所有组合。
为了彻底理解本发明，将在下列的描述中提出详细的结构，以便阐释本发明提出的技术方案。本发明的可选实施例详细描述如下，然而除了这些详细描述外，本发明还可以具有其他实施方式。
下面结合附图，对本申请的测距装置进行详细说明。在不冲突的情况下，下述的实施例及实施方式中的特征可以相互组合。
首先，参考图1和图2对本发明实施例中的一种测距装置的结构做更详细的示例性地描述，测距装置包括激光雷达，该测距装置仅作为示例，对于其他适合的测距装置也可以应用于本申请。
本发明各个实施例提供的方案可以应用于测距装置，该测距装置可以是激光雷达、激光测距设备等电子设备。在一种实施方式中，测距装置用于感测外部环境信息，例如，环境目标的距离信息、方位信息、反射强度信息、速度信息等。一种实现方式中，测距装置可以通过测量测距装置和探测物之间光传播的时间，即光飞行时间(Time-of-Flight，TOF)，来探测探测物到测距装置的距离。或者，测距装置也可以通过其他技术来探测探测物到测距装置的距离，例如基于相位移动(phase shift)测量的测距方法，或者基于频率移动(frequency shift)测量的测距方法，在此不做限制。
为了便于理解，以下将结合图1所示的测距装置100对测距的工作流程进行举例描述。
所述测距装置包括发射模块、接收模块和温度控制系统，所述发射模块用于出射光脉冲；所述接收模块用于接收经物体反射回的至少部分光脉冲，以及根据所述接收的至少部分光脉冲确定所述物体相对所述测距装置的距离。
具体地，如图1所示，所述发射模块包括发射电路110；所述接收模块包括接收电路120、采样电路130和运算电路140。
发射电路110可以出射光脉冲序列(例如激光脉冲序列)。接收电路120可以接收经过被探测物反射的光脉冲序列，并对该光脉冲序列进行光电转换，以得到电信号，再对电信号进行处理之后可以输出给采样电路130。采样电 路130可以对电信号进行采样，以获取采样结果。运算电路140可以基于采样电路130的采样结果，以确定测距装置100与被探测物之间的距离。
可选地，该测距装置100还可以包括控制电路150，该控制电路150可以实现对其他电路的控制，例如，可以控制各个电路的工作时间和/或对各个电路进行参数设置等。
应理解，虽然图1示出的测距装置中包括一个发射电路、一个接收电路、一个采样电路和一个运算电路，用于出射一路光束进行探测，但是本申请实施例并不限于此，发射电路、接收电路、采样电路、运算电路中的任一种电路的数量也可以是至少两个，用于沿相同方向或分别沿不同方向出射至少两路光束；其中，该至少两束光路可以是同时出射，也可以是分别在不同时刻出射。一个示例中，该至少两个发射电路中的发光芯片封装在同一个模块中。例如，每个发射电路包括一个激光发射芯片，该至少两个发射电路中的激光发射芯片中的die封装到一起，容置在同一个封装空间中。
一些实现方式中，除了图1所示的电路，测距装置100还可以包括扫描模块，用于将发射电路出射的至少一路光脉冲序列(例如激光脉冲序列)改变传播方向出射，以对视场进行扫描。示例性地，所述扫描模块在测距装置的视场内的扫描区域随着时间的累积而增加。
其中，可以将包括发射电路110、接收电路120、采样电路130和运算电路140的模块，或者，包括发射电路110、接收电路120、采样电路130、运算电路140和控制电路150的模块称为测距模块，该测距模块可以独立于其他模块，例如，扫描模块。
测距装置中可以采用同轴光路，也即测距装置出射的光束和经反射回来的光束在测距装置内共用至少部分光路。例如，发射电路出射的至少一路激光脉冲序列经扫描模块改变传播方向出射后，经探测物反射回来的激光脉冲序列经过扫描模块后入射至接收电路。或者，测距装置也可以采用异轴光路，也即测距装置出射的光束和经反射回来的光束在测距装置内分别沿不同的光路传输。图2示出了本发明的测距装置采用同轴光路的一种实施例的示意图。
测距装置200包括测距模块210，测距模块210包括发射器203(可以包括上述的发射电路)、准直元件204、探测器205(可以包括上述的接收电路、采样电路和运算电路)和光路改变元件206。测距模块210用于发射光束，且接收回光，将回光转换为电信号。其中，发射器203可以用于发射光脉冲 序列。在一个实施例中，发射器203可以发射激光脉冲序列。可选的，发射器203发射出的激光束为波长在可见光范围之外的窄带宽光束。准直元件204设置于发射器的出射光路上，用于准直从发射器203发出的光束，将发射器203发出的光束准直为平行光出射至扫描模块。准直元件还用于会聚经探测物反射的回光的至少一部分。该准直元件204可以是准直透镜或者是其他能够准直光束的元件。
在图2所示实施例中，通过光路改变元件206来将测距装置内的发射光路和接收光路在准直元件204之前合并，使得发射光路和接收光路可以共用同一个准直元件，使得光路更加紧凑。在其他的一些实现方式中，也可以是发射器203和探测器205分别使用各自的准直元件，将光路改变元件206设置在准直元件之后的光路上。
在图2所示实施例中，由于发射器203出射的光束的光束孔径较小，测距装置所接收到的回光的光束孔径较大，所以光路改变元件可以采用小面积的反射镜来将发射光路和接收光路合并。在其他的一些实现方式中，光路改变元件也可以采用带通孔的反射镜，其中该通孔用于透射发射器203的出射光，反射镜用于将回光反射至探测器205。这样可以减小采用小反射镜的情况中小反射镜的支架会对回光的遮挡。
在图2所示实施例中，光路改变元件偏离了准直元件204的光轴。在其他的一些实现方式中，光路改变元件也可以位于准直元件204的光轴上。
测距装置200还包括扫描模块202。扫描模块202放置于测距模块210的出射光路上，扫描模块202用于改变经准直元件204出射的准直光束219的传输方向并投射至外界环境，并将回光投射至准直元件204。回光经准直元件204汇聚到探测器205上。
在一个实施例中，扫描模块202可以包括至少一个光学元件，用于改变光束的传播路径，其中，该光学元件可以通过对光束进行反射、折射、衍射等等方式来改变光束传播路径，例如所述光学元件包括至少一个具有非平行的出射面和入射面的光折射元件。例如，扫描模块202包括透镜、反射镜、棱镜、振镜、光栅、液晶、光学相控阵(Optical Phased Array)或上述光学元件的任意组合。一个示例中，至少部分光学元件是运动的，例如通过驱动模块来驱动该至少部分光学元件进行运动，该运动的光学元件可以在不同时刻将光束反射、折射或衍射至不同的方向。在一些实施例中，扫描模块202的多个光学元件可以绕共同的轴209旋转或振动，每个旋转或振动的光学元件 用于不断改变入射光束的传播方向。在一个实施例中，扫描模块202的多个光学元件可以以不同的转速旋转，或以不同的速度振动。在另一个实施例中，扫描模块202的至少部分光学元件可以以基本相同的转速旋转。在一些实施例中，扫描模块的多个光学元件也可以是绕不同的轴旋转。在一些实施例中，扫描模块的多个光学元件也可以是以相同的方向旋转，或以不同的方向旋转；或者沿相同的方向振动，或者沿不同的方向振动，在此不作限制。
在一个实施例中，扫描模块202包括第一光学元件214和与第一光学元件214连接的驱动器216，驱动器216用于驱动第一光学元件214绕转动轴209转动，使第一光学元件214改变准直光束219的方向。第一光学元件214将准直光束219投射至不同的方向。在一个实施例中，准直光束219经第一光学元件改变后的方向与转动轴209的夹角随着第一光学元件214的转动而变化。在一个实施例中，第一光学元件214包括相对的非平行的一对表面，准直光束219穿过该对表面。在一个实施例中，第一光学元件214包括厚度沿至少一个径向变化的棱镜。在一个实施例中，第一光学元件214包括楔角棱镜，对准直光束219进行折射。
在一个实施例中，扫描模块202还包括第二光学元件215，第二光学元件215绕转动轴209转动，第二光学元件215的转动速度与第一光学元件214的转动速度不同。第二光学元件215用于改变第一光学元件214投射的光束的方向。在一个实施例中，第二光学元件215与另一驱动器217连接，驱动器217驱动第二光学元件215转动。第一光学元件214和第二光学元件215可以由相同或不同的驱动器驱动，使第一光学元件214和第二光学元件215的转速和/或转向不同，从而将准直光束219投射至外界空间不同的方向，可以扫描较大的空间范围。在一个实施例中，控制器218控制驱动器216和217，分别驱动第一光学元件214和第二光学元件215。第一光学元件214和第二光学元件215的转速可以根据实际应用中预期扫描的区域和样式确定。驱动器216和217可以包括电机或其他驱动器。
在一个实施例中，第二光学元件215包括相对的非平行的一对表面，光束穿过该对表面。在一个实施例中，第二光学元件215包括厚度沿至少一个径向变化的棱镜。在一个实施例中，第二光学元件215包括楔角棱镜。
一个实施例中，扫描模块202还包括第三光学元件(图未示)和用于驱动第三光学元件运动的驱动器。可选地，该第三光学元件包括相对的非平行的一对表面，光束穿过该对表面。在一个实施例中，第三光学元件包括厚度 沿至少一个径向变化的棱镜。在一个实施例中，第三光学元件包括楔角棱镜。第一、第二和第三光学元件中的至少两个光学元件以不同的转速和/或转向转动。
在一个实施例中，所述扫描模块包括在所述光脉冲序列的出射光路上依次排布的2个或3个所述光折射元件。可选地，所述扫描模块中的至少2个所述光折射元件在扫描过程中旋转，以改变所述光脉冲序列的方向。
所述扫描模块在至少部分不同时刻的扫描路径不同，扫描模块202中的各光学元件旋转可以将光投射至不同的方向，例如投射的光211的方向和方向213，如此对测距装置200周围的空间进行扫描。当扫描模块202投射出的光211打到探测物201时，一部分光被探测物201沿与投射的光211相反的方向反射至测距装置200。探测物201反射的回光212经过扫描模块202后入射至准直元件204。
探测器205与发射器203放置于准直元件204的同一侧，探测器205用于将穿过准直元件204的至少部分回光转换为电信号。
一个实施例中，各光学元件上镀有增透膜。可选的，增透膜的厚度与发射器203发射出的光束的波长相等或接近，能够增加透射光束的强度。
一个实施例中，测距装置中位于光束传播路径上的一个元件表面上镀有滤光层，或者在光束传播路径上设置有滤光器，用于至少透射发射器所出射的光束所在波段，反射其他波段，以减少环境光给接收器带来的噪音。
在一些实施例中，发射器203可以包括激光二极管，通过激光二极管发射纳秒级别的激光脉冲。进一步地，可以确定激光脉冲接收时间，例如，通过探测电信号脉冲的上升沿时间和/或下降沿时间确定激光脉冲接收时间。如此，测距装置200可以利用脉冲接收时间信息和脉冲发出时间信息计算TOF，从而确定探测物201到测距装置200的距离。测距装置200探测到的距离和方位可以用于遥感、避障、测绘、建模、导航等。
在一种实施方式中，本发明实施方式的测距装置可应用于移动平台，测距装置可安装在移动平台的平台本体。具有测距装置的移动平台可对外部环境进行测量，例如，测量移动平台与障碍物的距离用于避障等用途，和对外部环境进行二维或三维的测绘。在某些实施方式中，移动平台包括无人飞行器、汽车、遥控车、机器人、船、相机中的至少一种。当测距装置应用于无人飞行器时，平台本体为无人飞行器的机身。当测距装置应用于汽车时，平台本体为汽车的车身。该汽车可以是自动驾驶汽车或者半自动驾驶汽车，在 此不做限制。当测距装置应用于遥控车时，平台本体为遥控车的车身。当测距装置应用于机器人时，平台本体为机器人。当测距装置应用于相机时，平台本体为相机本身。
例如激光雷达的测距装置在扫描时，激光的发射角度不断变化，但是这些激光的发射角度在激光雷达的扫描视场内分布并不一定均匀，差异能有几倍到几百倍，甚至更多，扫描角度不均匀导致不同区域点云图像密度也不均匀，也即所述测距装置在扫描视场中具有非均匀的扫描密度，也即具有非规则的采样模式，因此导致测距装置获取的初始点云的均匀性差，如图3所示，其在大体中心区域的扫描密度高于其他区域，该区域的点云图像密度也高于其他区域。
例如激光雷达的测距装置输出的点云的均匀度、噪点等是点云质量的重要参数，影响着点云的显示效果、进一步处理等。目前改善点云质量，通常通过改善硬件采样以及点云生成算法来进行。然而上述方案对整个系统的要求较高，通常会顾此失彼，受到系统搭建难度、功耗、成本、扫描速率等的制约。在实际中，经优化设计的系统所输出的点云仍总是存在一些质量问题。
鉴于上述问题的存在，本发明提供一种点云重采样的方法，所述方法包括：对测距装置获取的初始点云进行重采样，以获得重采样后的重采样点云，其中，所述测距装置在扫描视场中具有非均匀的扫描密度，所述重采样点云的分布均匀性高于所述初始点云的分布均匀性，和/或，所述重采样点云的噪声低于所述初始点云的噪声。通过上述方法，可以将具有非规则采样模式的初始点云转换为均匀采样的重采样点云，可以增加或减少点云的采样密度。重采样后的点云，可以更适合后续处理，和/或拥有更好的显示效果，和/或隐藏特定的硬件采样模式等，以及所述重采样点云的噪声低于所述初始点云的噪声，因此提高了测距准确性。
下面结合附图，对本申请的点云重采样的方法和装置以及系统进行详细说明。在不冲突的情况下，下述的实施例及实施方式中的特征可以相互组合。
在一个实施例中，本发明的点云重采样的方法包括：对测距装置获取的初始点云进行重采样，以获得重采样后的重采样点云，其中，所述测距装置在扫描视场中具有非均匀的扫描密度，所述重采样点云的分布均匀性高于所述初始点云的分布均匀性，如图4所示，因此，重采样点云更适合后续处理，例如物体识别、图像融合等后续处理，拥有更好的显示效果以及隐藏特定的 硬件采样模式等优点。
在一个实施例中，通过在初始点云的点云间隙处进行插值后获得重采样点云的分布均匀性高于所述初始点云的分布均匀性。在一个示例中，所述重采样点云在参考面上的投影点相比所述初始点云在该参考面上的投影点的均匀度高。其中，所述参考面包括平面、球面、柱面中的至少一种，或者其他适合的参考面。
在一个实施例中，所述参考面包括平面，所述平面为与所述测距装置所发射的光脉冲序列的中心轴相垂直的面，其中，该平面也可以为测距装置的像平面，或者其他适合的平面，例如不与所述测距装置所发射的光脉冲序列的中心轴相垂直的面。可选地，初始点云或重采样点云在参考面上的投影点，可以指的是：以该参考面作为投影面，以测距装置(例如激光雷达)所在的位置作为中心点，将所述初始点云或所述重采样点云的点云点分别向该参考面进行透视投影得到各自位于该参考面上的投影点。所述重采样点云在参考面上的投影点相比所述初始点云在该参考面上的投影点的均匀度高。在另一个实施例中，所述参考面包括球面，例如，球面为以测距装置为中心的球面。在其他实施例中，所述参考面包括柱面，例如，所述柱面的中轴线为经过所述测距装置的竖直线，其中，该竖直线例如可以垂直测距装置的中心轴，或者，该竖直线例如可以垂直于水平面。
在一个实施例中，所述重采样点云在所述参考面上的至少部分投影点均匀分布，例如，在参考面中心区域的部分投影点均匀分布；或者，在其他部分区域的部分投影点均匀分布；或者，所述重采样点云在整个参考面上的投影点均匀分布；或者，所述重采样点云在所述参考面上的投影点中的至少部分投影点在该参考面的至少一条轨迹上等角度排布，如图10所示，其中，等角度排布是指至少部分投影点中，任意相邻两个投影点与测距装置的连线形成的夹角是相同的；或者，如图12所述重采样点云在所述参考面上的投影点中的至少部分投影点在该参考面的至少一条轨迹上等间隔排布，值得一提的是，等间隔排布例如相邻投影点之间的直线距离相同。
所述轨迹包括呈规则的直线状、规则的曲线状或者规则的折线状的轨迹或者其他适合的规则的轨迹。可选地，所述轨迹还可以包括沿水平方向和/或竖直方向上延伸的轨迹。在一个实施例中，如图10所示，轨迹包括呈规则的 曲线状的轨迹，并且该轨迹包括沿水平方向延伸的轨迹(大体为规则的曲线状)和沿竖直方向延伸的轨迹(大体为规则的直线状)。在另一实施例中，如图12所示，轨迹包括呈规则的直线状轨迹，并且，该轨迹包括沿水平方向和竖直方向延伸的轨迹。在又一个实施例中，所述轨迹包括多个同心环，同心环例如为同心圆环、同心方环、同心三角环、同心多边形环等等，其中，如图14所示，所述轨迹包括多个同心圆环。示例性地，该多个同心环的中心位于测距装置的中心轴上，以使轨迹对称的分布于该中心轴的周围，从而保证沿着该些轨迹均匀采样。在再一个实施例中，如图16所示，所述轨迹包括一条呈螺旋线的轨迹。在其他实施例中，所述轨迹包括多条沿一个中心点往外辐射的多条轨迹。使用例如非均匀扫描的激光雷达探测目标场景时，对应同一个物体的点云采样不连续性，不利于在点云展示时用户的友好度以及对点云的后续处理。
在一个实施例中，在所述初始点云的点云间隙处进行插值，以使重采样点云在参考面上的投影点相比所述初始点云在该参考面上的投影点的均匀度高。
上述实施例中主要针对重采样点云的特定排布模式进行了解释和说明，下面对点云重采样的方法的具体几个实施例进行描述。值得一提的是，测距装置探测的目标场景可以是室外场景也可以是室内场景，本申请实施例中主要以同一室内场景为例，例如，如图8所示的室内场景，该室内场景中具有位于前方的方板，位于左右两侧的桌子和座椅，位于屋顶的吊灯，以及地面等，图8为该室内场景的插值后的深度分布图，附图3、4、7、11、13、15、17均为该室内场景的点云图，其中，图3和图4中左图为初始点云图，图7为离散化后的点云投影图，图4中右图以及图13为一种重采样方法获得的重采样点云图，图11、图15和图17为其他不同类型重采样方法获得的重采样点云图。
可以采用任意适合的插值方法，在所述初始点云的点云间隙处进行插值。在一个示例中，可以先对初始点云进行降维处理，在低维度上对初始点云进行插值，以降低插值的难度。例如，将三维的初始点云投影到二维的参考面上，以获得位于参考面上的图像，例如，获得位于参考面上的点云投影图(或深度分布图)。如图8所示，在一个示例中，为了获得例如图8所示的图像， 所述插值的方法包括以下步骤A1至步骤A3：在步骤A1，将所述初始点云投影到参考面上，得到所述初始点云的投影点在参考面上的位置，以及参考点的数值(或者像素值)，该些投影点的数值包括其对应的初始点云点的特征信息，例如深度信息和/或反射率信息，例如，如图5所示，参考面(也即投影面)为平面，该平面为与所述测距装置所发射的光脉冲序列的中心轴相垂直的面，其中，该平面也可以为测距装置的像平面，或者其他适合的平面，可以以该参考面作为投影面，以测距装置(例如激光雷达)所在的位置作为中心点，将所述初始点云的点云点向该参考面进行透视投影得到位于该参考面上的投影点，例如图6所示的初始点云在参考面上的投影示意图，可见，在未进行重采样之前，初始点云的采样密度不均匀，中心区域的采样密度明显高于其他区域。
可选地，所述插值的方法还包括步骤A2，在所述参考面上将所述初始点云的投影点进行离散化，该步骤可以根据需要选择性进行。一个示例中，离散化包括采用一个像素取代落入到该像素内的一个或多个初始点云的投影点。可选的，该像素的像素值是根据落入到该像素中的初始点云的投影点计算的。以及步骤A2还能够方便确定插值的位置。
更具体地，所述离散化的方法包括：在所述参考面上将所述初始点云的投影点进行网格化，其中，插值点的位置包括至少部分空网格。一个像素可以包括一个或多个网格。在一个实施例中，所述网格是由所述参考面上多条相互垂直的轨迹形成的，例如，所述网格是由参考面上多条沿水平方向延伸和多条沿竖直方向延伸的交叉轨迹形成的，更进一步，所述网格可以是多条沿水平方向延伸和多条沿竖直方法延伸的交叉的直线状轨迹形成的方格。在另一个实施例中，所述网格是由参考面上的多条同心圆环与从该同心圆环的中心向外辐射的多条轨迹交叉形成的。在另一个实施例中，所述网格是由参考面上的一条螺旋线，与从该螺旋线的中心向外辐射的多条轨迹交叉形成的。可以根据需要合理的设定相邻的轨迹线之间的尺寸，从而设定处网格的物理尺寸。
可选地所述网格化的方法包括：将所述网格的深度值取代该网格内的多个投影点的深度值，其中该网格的深度值是根据落入到该网格中的多个投影点的深度值确定的。
在实际的应用中，其中每个网格可以对应一个像素，该网格的深度值也即像素的像素值，由于网格的物理尺寸的因素，可能存在多个投影点放在同一个网格中的情况，那么该网格的深度值可以根据落入到该网格中的多个投影点的深度值确定的，例如，可以根据落入到该网格中的多个投影点的最小深度值确定，或者，可以根据落入到该网格中的多个投影点的平均深度值。在仅包括唯一的投影点的网格中，网格的深度值也即为该投影点的深度值。
进一步，所述插值的方法还包括步骤A3，在所述参考面上的所述投影点的间隙处确定插值点。其中，可以基于任意适合的插值方法确定插值点的深度值，例如所述插值点的深度值是基于邻近的所述投影点的深度值确定的，更进一步，所述插值点的深度值是基于最近邻的所述投影点的深度值确定，或者基于邻近的投影点的最小深度值确定。
在一个示例中，确定所述插值点的方法还包括：确定所述空网格的位置是否天空位置，其中，待插值的所述空网格包括非天空位置的空网格。为了保证天空信息以及以天空为背景的物体的正确呈现，对天空位置的空网格和非天空位置的空网格赋予不同的值，例如，对所述天空位置的空网格赋予特殊值，所述特殊值为与具有投影点的网格的像素值不同的值，例如，所述特殊值为0，或者其他适合的值。可以根据测距装置探测天空位置时的特性以及天空位置自身的特性来对参考面中的天空位置进行判别，主要有以下两种：1)天空位置比较大，也即面积较大；2)测距装置发出探测信号进行探测时，在天空位置有发射信号没有接收信号。
在一个实施例中，对于非天空位置的空网格进行插值的方法包括：基于邻域插值算法对非天空位置的空网格进行插值，数字图像中的常规领域插值算法均可以应用于本发明，邻域插值算法包括最近邻插值算法或线性插值算法等对非天空位置的空网格赋予适合的深度值，如图8所示为插值后获得的深度分布图，由此获得了一幅采样均匀的图像。
可选地，可以在该图像的基础上将插值后的点云进行任意模式的点云重采样，以获得具有特定排布模式的所述重采样点云，进而提供更多的点云呈现结果，该步骤可以根据需要选择性进行。可选地，可以在二维的参考面上进行点云重采样。一个示例中，修改投影点在该二维的参考面上位置，使得重采样投影点按特定模式排布。其中，修改后的投影点(也即重采样后的投 影点)的像素值可以根据重采样前的参考面上相同位置处的投影点的像素值确定，或者根据重采样前的参考面上相同位置处的邻近投影点的像素值确定，例如根据重采样前的参考面上相同位置处的邻近投影点的像素值插值得到，其中，本文中的像素值包括深度信息和/或反射率信息。下面对几种重采样的方法进行举例描述。
在图9-图11所示的一个具体实施例的点云重采样方法中，如图9所示，对例如激光雷达的测距装置的扫描方向(θ、
Figure PCTCN2019074585-appb-000001
)分别进行均匀采样，其中θ和
Figure PCTCN2019074585-appb-000002
对应为在如图9所示的极坐标系中的坐标，均匀采样例如均匀改变θ和
Figure PCTCN2019074585-appb-000003
然后对每一个重采样方向计算该重采样方向与参考面上的图像(例如图8所示的图像)的交点，通过例如前述的插值的方式获得该交点的取值，由此可以得到该重采样方向的像素值，进而确定一个重采样点云点。特别地，可以采用最近邻插值的方法，将该交点所在像素(也即网格)的取值作为该采样方向的像素值。基于上述方法，可以获得如图11所示的重采样点云，而所述重采样点云在所述参考面上的投影点中的至少部分投影点在该参考面的至少一条轨迹上等角度排布，如图10所示，该投影点对应于重采样方向集和该参考面上的图像的交点。
在图12-图13所示的另一个具体实施例的点云重采样方法中，该点云重采样方法也可以称为平面均匀重采样方法，例如，遍历前述插值后的图像中(也即参考面上的图像)的每一个像素，若该像素的像素值非零，则产生一个重采样点，该重采样点可以位于原点(例如以测距装置所在的位置为原点)和图像像素中心连线上，深度值为对应的像素的像素值，以此获得重采样点云，如图13所示，所述重采样点云的分布均匀性高于所述初始点云的分布均匀性。该重采样点云在所述参考面上的投影点中的至少部分投影点在该参考面的至少一条轨迹上等间隔排布。图12所示的为重采样点云在参考面上的投影点(该投影点对应于图像像素中心)，所述重采样点云在参考面上的投影点相比所述初始点云在该参考面上的投影点的均匀度高，如此，可获得一个平面均匀重采样点云。更具体地，还可以采用例如前述实施例的方式，先生成待采样的平面均匀点，该些均匀点等间隔排列，然后计算该些均匀点和参考面上的图像的交点，再利用插值的方式，来获得重采样点云。
在如图14至图17所示的再一个具体实施例的点云重采样方法中，按照 特定模式重采样，首先根据目标模式需求，生成一个重采样方向集，该些重采样方向集在参考面内沿着轨迹排布，例如图14所示，该轨迹包括多个同心圆环(也即均匀圆周采样)，或者，如图16所示，所述轨迹包括一条呈螺旋线的轨迹。对其中的任意一个重采样方向计算其与图像平面(也即参考面)的交点，进而基于该交点的取值，获得该重采样方向的重采样点云点的取值，从而获得如图15所示的重采样点云(基于均匀圆周采样)或如图17所示的重采样点云(基于螺旋采样)。
通过上述重采样方法获得重采样点云的分布均匀性高于所述初始点云的分布均匀性，可以将具有非规则采样模式的初始点云转换为均匀采样的重采样点云，重采样后的点云，可以更适合后续处理(例如物体识别、图像融合等)，和/或拥有更好的显示效果，和/或隐藏特定的硬件采样模式等。
在一个实施例中，所述重采样点云相比所述初始点云增加了多个位于同一个所述参考面上的点云点，也即重采样点云相比所述初始点云增加了多个位于同一个所述参考面上的投影点。进一步地，所述方法还可以包括：在所述初始点云中增加确定的插值点。再一个具体示例中，在初始点云采样方向集中添加新的重采样方向，以增加采样密度，其中，每个重采样方向对应的点云点作为重采样点云点，可以通过例如插值的方法计算该些重采样点云点的像素值。插值的方法可以基于前述实施例中的插值方法，例如最近邻插值算法或者线性插值算法等。通过增加点云密度重采样的方法，初始点云中增加确定的插值点，以形成新的高密度重采样点云，将具有非规则采样模式的初始点云转换为均有采样的重采样点云，采样后的重采样点云更适合后续处理，拥有更好的显示效果，隐藏特定的硬件采样模式等。
在本发明的另一个实施例中，所述初始点云不同区域的点云密度不同，对所述初始点云进行重采样包括：在点云密度高的区域对所述初始点云进行降采样。更具体地，所述测距装置在扫描视场的中心区域的扫描密度高于其余区域的扫描密度，对所述初始点云进行重采样包括：对中心区域的初始点云进行降采样，以均衡采样密度。通过均衡将采样的方法，可以减少点云密度高的区域的采样密度，将具有非规则采样模式的初始点云转换为均有采样的重采样点云，采样后的重采样点云更适合后续处理，拥有更好的显示效果，隐藏特定的硬件采样模式等。
在一个示例中，所述降采样的方法包括：随机或者均匀舍弃所述中心区域的部分网格对应的全部初始点云点，其中，所述网格是由在参考面上将所述初始点云的投影点进行网格化而形成的。在另一个实施例中，所述降采样的方法包括：限制落入所述中心区域内的至少部分网格内的初始点云点的数量低于阈值数量，其中，所述网格是由在参考面上将初始点云的投影点进行网格化而形成的。进一步，所述限制落入所述中心区域内的至少部分网格内的初始点云点的数量低于阈值数量的方法，包括：随机或者均匀舍弃所述至少部分网格内的部分初始点云点(也即舍弃与该部分初始点云点在该部分网格内的投影点)，直至该至少部分网格内投影点的个数在阈值之内。所述阈值数量根据实际的需要进行合理的设定，例如将中心区域内的所有网格中的投影点的数量除以中心区域的网格的数量获得平均值，将阈值数量设置为等于该平均值，控制中心区域内的至少部分网格中的投影点的数量低于该阈值，或者，将参考面上的所有初始点云的投影点的数量除以网格的总数量获得一个平均值，将阈值数量设置为等于该平均值，上述两种阈值的设置方式仅作为示例，对于其他能够起到均衡采样密度高的区域的点云数量的阈值也同样可以适用于本发明。
在本发明的再一个实施例中，对所述初始点云进行重采样包括：对所述初始点云进行降噪重采样，以使所述重采样点云的噪声低于所述初始点云的噪声。
在一个示例中，所述降噪重采样的方法包括：对所述初始点云在参考面上的图像进行滤噪处理。该处理应具有保边性，如采用双边滤波器等进行滤噪处理，经滤噪后的图像连续性较好、易于观察。
所述图像可以包括深度分布图或者反射率分布图，基于前述实施例中的方法获得该图像，具体地方法过程可以参考前述实施例的描述，本实施例中主要以所述图像为深度分布图的情况为例对重采样的方法进行说明，在一个示例中，获得所述图像的方法包括：将所述初始点云投影到参考面上，得到所述初始点云的投影点，该些投影点包括其对应的初始点云点的特征信息，例如深度信息和/或反射率信息，如图5所示；随后，在所述参考面上将所述初始点云的投影点进行网格化，其中，每个网格对应一个像素，该网格化的过程参考前述实施例中的描述，在此不做赘述；对部分空网格进行插值。示 例性地，根据落入到网格中的多个投影点的像素值确定该网格的像素值，例如网格的像素值取落入该网格的投影点的最小像素值；以及基于邻近网格的像素值确定空网格的像素值。
在一个示例中，所述基于邻近网格的像素值确定空网格的像素值的方法，具体为：确定所述空网格的位置是否天空位置，其中，基于邻近网格的像素值确定非天空位置的所述空网格的像素值。具体地可以参考前述实施例的描述，在此不做具体赘述，通过上述方法可以获得例如图8所示的深度分布图。
在一个示例中，所述降噪重采样的方法还包括：遍历所述初始点云的每个初始点云点；将每个所述初始点云点映射至所述滤噪处理后的所述图像；将每个所述初始点云点的像素值调整为其对应像素的像素值，或者，调整为所述初始点云点在所述图像的投影位置的相邻像素的像素值，在相邻像素的像素值还可能是插值后的像素值。在另一个示例中，所述降噪重采样的方法还包括：基于所述初始点云点调整后的像素值与该初始点云点的初始像素值进行阈值处理，在所述初始像素值和所述调整后的像素值之差大于目标阈值时，则表明该初始点云点很可能为噪点，因此舍弃该初始点云点。
经上述降噪处理，可以滤除两类因探测精度不佳所致的噪点，包括边缘上的噪点、初始点云中可能在同一采样方向上有多个点云点(其中深度较大者基本可认为是噪声)以及使平面表面毛糙化的点，进而使得例如平面的表面更加平整，物体的边缘更加清晰，便于后续处理的应用，提高测距装置的测量准确性。
在本发明另一个实施例中还提供一种点云重采样的方法，该实施例中的相关技术特征的描述可以参考前述实施例，下面主要对该实施例的方法中包括的步骤进行解释和说明，所述点云重采样的方法主要包括以下步骤：
首先，将初始点云投影到参考面上，得到所述初始点云的投影点，其中，所述初始点云的描述参考前述实施例，该参考面包括平面、球面、柱面中的至少一种，其中，本实施例中主要以参考面包括二维平面为例进行解释和说明；接着，在所述参考面上的所述投影点的间隙处确定插值点，以获得位于所述参考面上的图像，该图像可以是位于二维平面上的二维图像，例如二维深度分布图等，随后，基于所述图像进行重采样，以获得具有特定排布模式的重采样点云，或者，基于所述图像进行降噪重采样，以使所述重采样点云 的噪声低于所述初始点云的噪声，该降噪重采样具体方法参考前述实施例中的描述。
可选地，在投影之后确定插值点之前，还可以在所述参考面上将所述初始点云的投影点进行离散化，该离散化的具体步骤可以参考前述实施例，在此不做赘述。
在本实施例中，特定排布模式的重采样点云可以参考上文中对重采样点云的描述，例如是前述实施例中图11、图13、图15和图17中的任意一种排布模式的重采样点云。
基于本实施例的方法获得的所述重采样点云的分布均匀性高于所述初始点云的分布均匀性，和/或，所述重采样点云的噪声低于所述初始点云的噪声。通过上述方法，可以将具有非规则采样模式的初始点云转换为均匀采样的重采样点云，可以增加或减少点云的采样密度。重采样后的点云，可以更适合后续处理，和/或拥有更好的显示效果，和/或隐藏特定的硬件采样模式等，以及所述重采样点云的噪声低于所述初始点云的噪声，因此提高了测距准确性。
为了实现前述实施例中的方法，前述实施例中的技术特征同样适用于本实施例中，如图18所示，本发明实施例还提供一种点云重采样装置500，所述点云重采样装置500包括重采样模块510，用于对测距装置获取的初始点云进行重采样，以获得重采样后的重采样点云，其中，所述测距装置在扫描视场中具有非均匀的扫描密度，所述重采样点云的分布均匀性高于所述初始点云的分布均匀性，和/或，所述重采样点云的噪声低于所述初始点云的噪声。
在一个示例中，所述重采样点云在参考面上的投影点相比所述初始点云在该参考面上的投影点的均匀度高。所述参考面包括平面、球面、柱面中的至少一种。其中，所述平面为与所述测距装置所发射的光脉冲序列的中心轴相垂直的面；或者，所述球面为以所述测距装置为中心的球面；或者，所述柱面的中轴线为经过所述测距装置的竖直线。
进一步，所述重采样点云在所述参考面上的至少部分投影点均匀分布。
在一个实施例中，所述重采样点云在所述参考面上的投影点中的至少部分投影点在该参考面的至少一条轨迹上等间隔排布或者等角度排布。其中，所述轨迹包括呈规则的直线状、规则的曲线状或者规则的折线状的轨迹，或者其他适合的轨迹。在另一示例中，所述轨迹包括沿水平方向和/或竖直方向 上延伸的轨迹。
在一个示例中，所述轨迹包括多个同心环，该多个同心环之间可以具有相同的间隔或者不同的间隔。其中，所述多个同心环的中心位于所述测距装置的中心轴上。在另一个示例中，所述轨迹包括一条呈螺旋线的轨迹。在其他示例中，所述轨迹包括多条沿一个中心点往外辐射的多条轨迹。
在一个示例中，所述重采样点云相比所述初始点云增加了多个位于同一个所述参考面上的点云点。因此重采样点云均匀性更高。
所述点云重采样装置还包括图像生成模块，所述图像生成模块包括投影模块和插值模块，投影模块用于将所述初始点云投影到参考面上，得到所述初始点云的投影点；插值模块用于在所述初始点云的点云间隙处进行插值，更具体地，所述插值模块用于在所述参考面上的所述投影点的间隙处确定插值点，以获得位于所述参考面上的图像，可选地，所述插值点的取值是基于邻近的所述投影点的像素值(例如深度值)确定的，或者其他适合的插值方式也可以适用于本发明。
可选地，所述图像生成模块还包括离散化模块，用于在确定所述插值点之前在所述参考面上将所述初始点云的投影点进行离散化。离散化模块具体用于：在所述参考面上将所述初始点云的投影点进行网格化，其中，所述插值点的位置包括至少部分空网格。将所述网格的像素值取代该网格内的多个投影点的像素值，其中该网格的像素值是根据落入到该网格中的多个投影点的像素值确定的，则所述插值点的取值可以是基于邻近网格的像素值确定的。
可选地，所述网格是由所述参考面上多条相互垂直的轨迹形成的，例如，所述网格是由参考面上多条沿水平方向延伸和多条沿竖直方向延伸的交叉轨迹形成的。可选地，所述网格还可以是由参考面上的多条同心圆环与从该同心环的中心向外辐射的多条轨迹交叉形成的。或者，所述网格还可以是由参考面上的一条螺旋线，与从该螺旋线的中心向外辐射的多条轨迹交叉形成的。
所述插值模块还用于确定所述空网格的位置是否天空位置，其中，待插值的所述空网格包括非天空位置的空网格。为了保证天空信息以及以天空为背景的物体的正确呈现，对天空位置的空网格和非天空位置的空网格赋予不同的值，例如，对所述天空位置的空网格赋予特殊值，所述特殊值为与具有投影点的网格的像素值不同的值，例如，所述特殊值为0，或者其他适合的 值。可以根据测距装置探测天空位置时的特性以及天空位置自身的特性来对参考面中的天空位置进行判别，主要有以下两种：1)天空位置比较大，也即面积较大；2)测距装置发出探测信号进行探测时，在天空位置有发射信号没有接收信号。
在一个实施例中，对于非天空位置的空网格进行插值的方法包括：基于邻域插值算法对非天空位置的空网格进行插值，数字图像中的常规领域插值算法均可以应用于本发明，邻域插值算法包括最近邻插值算法或线性插值算法等对非天空位置的空网格赋予适合的像素值，由此获得了一幅采样均匀的图像，所述重采样模块还具体用于：基于所述图像进行重采样，以获得具有特定排布模式的所述重采样点云。
其中，所述重采样模块具体用于基于所述图像进行重采样，以获得具有特定排布模式的所述重采样点云的过程可以参考前述实施例中的描述，在此不做赘述。
在另一个示例中，所述重采样模块还用于：在所述初始点云中增加确定的插值点。所述重采样点云相比所述初始点云增加了多个位于同一个所述参考面上的点云点，也即重采样点云相比所述初始点云增加了多个位于同一个所述参考面上的投影点。再一个具体示例中，在初始点云采样方向集中添加新的重采样方向，以增加采样密度，其中，每个重采样方向对应的点云点作为重采样点云点，可以通过例如插值的方法计算该些重采样点云点的像素值(例如深度值)。插值的方法可以基于前述实施例中的插值方法，例如最近邻插值算法或者线性插值算法等。通过增加点云密度重采样的方法，初始点云中增加确定的插值点，以形成新的高密度重采样点云，将具有非规则采样模式的初始点云转换为均有采样的重采样点云，采样后的重采样点云更适合后续处理，拥有更好的显示效果，隐藏特定的硬件采样模式等。
在另一个实施例中，所述初始点云不同区域的点云密度不同，所述重采样模块还包括降采样模块，用于在点云密度高的区域对所述初始点云进行降采样。
在一个示例中，所述测距装置在扫描视场的中心区域的扫描密度高于其余区域的扫描密度，降采样模块具体用于对中心区域的初始点云进行降采样。
在另一个示例中，所述降采样模块具体用于：随机或者均匀舍弃所述中 心区域的部分网格对应的全部初始点云点，其中，所述网格是由在参考面上将所述初始点云的投影点进行网格化而形成的。
在再一个示例中，所述降采样模块具体用于：限制落入所述中心区域内的至少部分网格内的初始点云点的数量低于阈值数量，其中，所述网格是由在参考面上将初始点云的投影点进行网格化而形成的。在其他示例中，所述降采样模块具体还用于：随机或者均匀舍弃所述至少部分网格内的部分初始点云点。
在再一个实施例中，所述重采样模块还包括降噪重采样模块，用于对所述初始点云进行降噪重采样，以使所述重采样点云的噪声低于所述初始点云的噪声。示例性地，所述降噪重采样模块具体用于：对所述初始点云在参考面上的图像进行滤噪处理。
在一个示例中，所述降噪重采样模块具体还用于：遍历所述初始点云的每个初始点云点；将每个所述初始点云点映射至所述滤噪处理后的所述图像；将每个所述初始点云点的像素值(例如深度值)调整为其对应像素的像素值，或者，调整为所述初始点云点在所述图像的投影位置的相邻像素的像素值。其中，所述降噪重采样模块进一步用于：基于所述初始点云点调整后的像素值与该初始点云点的初始像素值进行阈值处理，在所述初始像素值和所述调整后的像素值之差大于目标阈值时，舍弃该初始点云点。其中，所述图像包括深度分布图，或者其他适合的图像，例如反射率分布图等。
在一个示例中，点云重采样装置，还包括用于获得所述图像的图像生成模块，所述图像生成模块包括离散化模块和插值模块，离散化模块用于在所述参考面上将所述初始点云的投影点进行网格化，其中，每个网格对应一个像素，所述离散化模块还用于根据落入到网格中的多个投影点的像素值确定该网格的像素值；以及；插值模块用于对部分空网格进行插值，更具体地，所述插值模块用于基于邻近网格的像素值确定空网格的像素值。
所述插值模块具体还用于：确定所述空网格的位置是否天空位置，其中，基于邻近网格的像素值确定非天空位置的所述空网格的像素值。
示例性地，根据本发明实施例的点云重采样方法可以在包括存储器和处理器的设备、装置或者系统中实现。
图19示出了本发明一个实施例中的点云重采样系统600的示意性框图。
点云重采样系统600包括一个或多个测距装置610，一个或多个处理器630、一个或多个存储装置620。可选地，点云重采样系统600还可以包括输入装置(未示出)、输出装置(未示出)以及图像传感器(未示出)中的至少一个，这些组件通过总线系统和/或其它形式的连接机构(未示出)互连。应当注意，图19所示的点云重采样系统600的组件和结构只是示例性的，而非限制性的，根据需要，所述点云重采样系统600也可以具有其他组件和结构，例如还可以包括用于收发信号的收发器。
测距装置610的具体结构参考前述实施例中的描述在此不做赘述。
所述存储装置620也即存储器用于存储处理器可执行指令的存储器，例如用于存在用于实现根据本发明实施例的点云重采样的方法中的相应步骤和程序指令。可以包括一个或多个计算机程序产品，所述计算机程序产品可以包括各种形式的计算机可读存储介质，例如易失性存储器和/或非易失性存储器。所述易失性存储器例如可以包括随机存取存储器(RAM)和/或高速缓冲存储器(cache)等。所述非易失性存储器例如可以包括只读存储器(ROM)、硬盘、闪存等。
所述输入装置可以是用户用来输入指令的装置，并且可以包括键盘、鼠标、麦克风和触摸屏等中的一个或多个。
所述输出装置可以向外部(例如用户)输出各种信息(例如图像或声音)，并且可以包括显示器、扬声器等中的一个或多个。
通信接口(未示出)用于点云重采样系统600和其他设备之间进行通信，包括有线或者无线方式的通信。点云重采样系统600可以接入基于通信标准的无线网络，如WiFi、2G、3G、4G、5G或它们的组合。在一个示例性实施例中，通信接口经由广播信道接收来自外部广播管理系统的广播信号或广播相关信息。在一个示例性实施例中，所述通信接口还包括近场通信(NFC)模块，以促进短程通信。例如，在NFC模块可基于射频识别(RFID)技术，红外数据协会(IrDA)技术，超宽带(UWB)技术，蓝牙(BT)技术和其他技术来实现。
所述处理器630可以是中央处理单元(CPU)、图像处理单元(GPU)、专用集成电路(ASIC)、现场可编程门阵列(FPGA)或者具有数据处理能力和/或指令执行能力的其它形式的处理单元，并且可以控制所述点云重采样系统600中的其它组件以执行期望的功能。所述处理器能够执行所述存储装置 620中存储的所述指令，以执行本文描述的点云重采样的方法。例如，处理器630能够包括一个或多个嵌入式处理器、处理器核心、微型处理器、逻辑电路、硬件有限状态机(FSM)、数字信号处理器(DSP)或它们的组合。
在所述计算机可读存储介质上可以存储一个或多个计算机程序指令，处理器630可以运行存储装置620存储的所述程序指令，以实现本文所述的本发明实施例中(由处理器实现)的功能以及/或者其它期望的功能，例如以执行根据本发明实施例的点云重采样的方法的相应步骤，并且用于实现根据本发明实施例中的点云重采样装置中的各个模块。在所述计算机可读存储介质中还可以存储各种应用程序和各种数据，例如所述应用程序使用和/或产生的各种数据等。
另外，本发明实施例还提供了一种计算机存储介质，其上存储有计算机程序。当所述计算机程序由处理器执行时，可以实现本发明实施例的基于点云数据的点云重采样方法的各个步骤或前述点云重采样装置中的各组成模块。例如，所述计算机存储介质例如可以包括智能电话的存储卡、平板电脑的存储部件、个人计算机的硬盘、只读存储器(ROM)、可擦除可编程只读存储器(EPROM)、便携式紧致盘只读存储器(CD-ROM)、USB存储器、或者上述存储介质的任意组合。所述计算机可读存储介质可以是一个或多个计算机可读存储介质的任意组合。例如一个计算机可读存储介质包含用于将所述点云数据转换为二维图像的计算机可读的程序代码，另一个计算机可读存储介质包含用于对所述二维图像进行物体分割的计算机可读的程序代码等。
尽管这里已经参考附图描述了示例实施例，应理解上述示例实施例仅仅是示例性的，并且不意图将本发明的范围限制于此。本领域普通技术人员可以在其中进行各种改变和修改，而不偏离本发明的范围和精神。所有这些改变和修改意在被包括在所附权利要求所要求的本发明的范围之内。
本领域普通技术人员可以意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本发明的范围。
在本申请所提供的几个实施例中，应该理解到，所揭露的设备和方法，可以通过其它的方式实现。例如，以上所描述的设备实施例仅仅是示意性的， 例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个设备，或一些特征可以忽略，或不执行。
在此处所提供的说明书中，说明了大量具体细节。然而，能够理解，本发明的实施例可以在没有这些具体细节的情况下实践。在一些实例中，并未详细示出公知的方法、结构和技术，以便不模糊对本说明书的理解。
类似地，应当理解，为了精简本发明并帮助理解各个发明方面中的一个或多个，在对本发明的示例性实施例的描述中，本发明的各个特征有时被一起分组到单个实施例、图、或者对其的描述中。然而，并不应将该本发明的方法解释成反映如下意图：即所要求保护的本发明要求比在每个权利要求中所明确记载的特征更多的特征。更确切地说，如相应的权利要求书所反映的那样，其发明点在于可以用少于某个公开的单个实施例的所有特征的特征来解决相应的技术问题。因此，遵循具体实施方式的权利要求书由此明确地并入该具体实施方式，其中每个权利要求本身都作为本发明的单独实施例。
本领域的技术人员可以理解，除了特征之间相互排斥之外，可以采用任何组合对本说明书(包括伴随的权利要求、摘要和附图)中公开的所有特征以及如此公开的任何方法或者设备的所有过程或单元进行组合。除非另外明确陈述，本说明书(包括伴随的权利要求、摘要和附图)中公开的每个特征可以由提供相同、等同或相似目的替代特征来代替。
此外，本领域的技术人员能够理解，尽管在此所述的一些实施例包括其它实施例中所包括的某些特征而不是其它特征，但是不同实施例的特征的组合意味着处于本发明的范围之内并且形成不同的实施例。例如，在权利要求书中，所要求保护的实施例的任意之一都可以以任意的组合方式来使用。
本发明的各个部件实施例可以以硬件实现，或者以在一个或者多个处理器上运行的软件模块实现，或者以它们的组合实现。本领域的技术人员应当理解，可以在实践中使用微处理器或者数字信号处理器(DSP)来实现根据本发明实施例的一些模块的一些或者全部功能。本发明还可以实现为用于执行这里所描述的方法的一部分或者全部的装置程序(例如，计算机程序和计算机程序产品)。这样的实现本发明的程序可以存储在计算机可读介质上，或者可以具有一个或者多个信号的形式。这样的信号可以从因特网网站上下载得到，或者在载体信号上提供，或者以任何其他形式提供。
应该注意的是上述实施例对本发明进行说明而不是对本发明进行限制， 并且本领域技术人员在不脱离所附权利要求的范围的情况下可设计出替换实施例。在权利要求中，不应将位于括号之间的任何参考符号构造成对权利要求的限制。本发明可以借助于包括有若干不同元件的硬件以及借助于适当编程的计算机来实现。在列举了若干装置的单元权利要求中，这些装置中的若干个可以是通过同一个硬件项来具体体现。单词第一、第二、以及第三等的使用不表示任何顺序。可将这些单词解释为名称。

Claims (91)

  1. 一种点云重采样的方法，其特征在于，所述方法包括：
    对测距装置获取的初始点云进行重采样，以获得重采样后的重采样点云，其中，所述测距装置在扫描视场中具有非均匀的扫描密度，所述重采样点云的分布均匀性高于所述初始点云的分布均匀性，和/或，所述重采样点云的噪声低于所述初始点云的噪声。
  2. 如权利要求1所述的方法，其特征在于，所述对测距装置输出的初始点云进行重采样包括：
    在所述初始点云的点云间隙处进行插值。
  3. 如权利要求2所述的方法，其特征在于，所述重采样点云在参考面上的投影点相比所述初始点云在该参考面上的投影点的均匀度高。
  4. 如权利要求3所述的方法，其特征在于，所述参考面包括平面、球面、柱面中的至少一种。
  5. 如权利要求4所述的方法，其特征在于，
    所述平面为与所述测距装置所发射的光脉冲序列的中心轴相垂直的面；或者，
    所述球面为以所述测距装置为中心的球面；或者，
    所述柱面的中轴线为经过所述测距装置的竖直线。
  6. 如权利要求3所述的方法，其特征在于，所述重采样点云在所述参考面上的至少部分投影点均匀分布。
  7. 如权利要求6所述的方法，其特征在于，所述重采样点云在所述参考面上的投影点中的至少部分投影点在该参考面的至少一条轨迹上等间隔排布或者等角度排布。
  8. 如权利要求7所述的方法，其特征在于，所述轨迹包括呈规则的直线状、规则的曲线状或者规则的折线状的轨迹。
  9. 如权利要求7所述的方法，其特征在于，所述轨迹包括沿水平方向和/或竖直方向上延伸的轨迹。
  10. 如权利要求7所述的方法，其特征在于，所述轨迹包括多个同心环。
  11. 如权利要求10所述的方法，其特征在于，所述多个同心环的中心位于所述测距装置的中心轴上。
  12. 如权利要求7所述的方法，其特征在于，所述轨迹包括一条呈螺旋线的轨迹。
  13. 如权利要求7所述的方法，其特征在于，所述轨迹包括多条沿一个中心点往外辐射的多条轨迹。
  14. 如权利要求3所述的方法，其特征在于，所述重采样点云相比所述初始点云增加了多个位于同一个所述参考面上的点云点。
  15. 如权利要求3所述的方法，其特征在于，所述插值的方法包括：将所述初始点云投影到参考面上，得到所述初始点云的投影点；
    在所述参考面上的所述投影点的间隙处确定插值点，以获得位于所述参考面上的图像。
  16. 如权利要求15所述的方法，其特征在于，所述插值点的深度值是基于邻近的所述投影点的深度值确定的。
  17. 如权利要求15所述的方法，其特征在于，所述插值的方法还包括：在确定所述插值点之前，在所述参考面上将所述初始点云的投影点进行离散化。
  18. 如权利要求17所述的方法，其特征在于，所述离散化的方法包括：
    在所述参考面上将所述初始点云的投影点进行网格化，其中，所述插值点的位置包括至少部分空网格。
  19. 如权利要求15-18任一项所述的方法，其特征在于，所述对测距装置获取的初始点云进行重采样包括：
    基于所述图像进行重采样，以获得具有特定排布模式的所述重采样点云。
  20. 如权利要求18所述的方法，其特征在于，所述网格化的方法包括：
    将所述网格的深度值取代该网格内的多个投影点的深度值，其中该网格的深度值是根据落入到该网格中的多个投影点的深度值确定的；以及
    所述插值点的深度值是基于邻近网格的深度值确定的。
  21. 如权利要求18所述的方法，其特征在于，确定所述插值点的方法还包括：
    确定所述空网格的位置是否天空位置，其中，待插值的所述空网格包括非天空位置的空网格。
  22. 如权利要求18所述的方法，其特征在于，所述网格是由所述参考面 上多条相互垂直的轨迹形成的。
  23. 如权利要求22所述的方法，其特征在于，所述网格是由参考面上多条沿水平方向延伸和多条沿竖直方向延伸的交叉轨迹形成的。
  24. 如权利要求18所述的方法，其特征在于，所述网格是由参考面上的多条同心圆环与从该同心环的中心向外辐射的多条轨迹交叉形成的。
  25. 如权利要求18所述的方法，其特征在于，所述网格是由参考面上的一条螺旋线，与从该螺旋线的中心向外辐射的多条轨迹交叉形成的。
  26. 如权利要求16至18、20至25任一项所述的方法，其特征在于，所述方法还包括：
    在所述初始点云中增加确定的插值点。
  27. 如权利要求1所述的方法，其特征在于，所述初始点云不同区域的点云密度不同，对所述初始点云进行重采样包括：
    在点云密度高的区域对所述初始点云进行降采样。
  28. 如权利要求27所述的方法，其特征在于，所述测距装置在扫描视场的中心区域的扫描密度高于其余区域的扫描密度，对所述初始点云进行重采样包括：对中心区域的初始点云进行降采样。
  29. 如权利要求28所述的方法，其特征在于，所述降采样的方法包括：
    随机或者均匀舍弃所述中心区域的部分网格对应的全部初始点云点，其中，所述网格是由在参考面上将所述初始点云的投影点进行网格化而形成的。
  30. 如权利要求28所述的方法，其特征在于，所述降采样的方法包括：
    限制落入所述中心区域内的至少部分网格内的初始点云点的数量低于阈值数量，其中，所述网格是由在参考面上将初始点云的投影点进行网格化而形成的。
  31. 如权利要求30所述的方法，其特征在于，所述限制落入所述中心区域内的至少部分网格内的初始点云点的数量低于阈值数量的方法，包括：
    随机或者均匀舍弃所述至少部分网格内的部分初始点云点。
  32. 如权利要求1所述的方法，其特征在于，对所述初始点云进行重采样包括：对所述初始点云进行降噪重采样，以使所述重采样点云的噪声低于所述初始点云的噪声。
  33. 如权利要求32所述的方法，其特征在于，所述降噪重采样的方法包 括：
    对所述初始点云在参考面上的图像进行滤噪处理。
  34. 如权利要求33所述的方法，其特征在于，所述降噪重采样的方法还包括：
    遍历所述初始点云的每个初始点云点；
    将每个所述初始点云点映射至所述滤噪处理后的所述图像；
    将每个所述初始点云点的深度值调整为其对应像素的深度值，或者，调整为所述初始点云点在所述图像的投影位置的相邻像素的深度值。
  35. 如权利要求34所述的方法，其特征在于，所述降噪重采样的方法还包括：
    基于所述初始点云点调整后的深度值与该初始点云点的初始深度值进行阈值处理，在所述初始深度值和所述调整后的深度值之差大于目标阈值时，舍弃该初始点云点。
  36. 如权利要求32至35任一项所述的方法，其特征在于，所述图像包括深度分布图。
  37. 如权利要求36所述的方法，其特征在于，获得所述图像的方法包括：
    在所述参考面上将所述初始点云的投影点进行网格化，其中，每个网格对应一个像素；
    对部分空网格进行插值。
  38. 如权利要求36所述的方法，其特征在于，获得所述图像的方法包括：
    根据落入到网格中的多个投影点的深度值确定该网格的深度值；以及
    基于邻近网格的深度值确定空网格的深度值。
  39. 如权利要求38所述的方法，其特征在于，所述基于邻近网格的深度值确定空网格的深度值的方法，具体为：
    确定所述空网格的位置是否天空位置，其中，基于邻近网格的深度值确定非天空位置的所述空网格的深度值。
  40. 如权利要求1所述的方法，其特征在于，所述测距装置包括用于出射光脉冲序列的发射模块，以及用于改变光脉冲序列的方向以对视场进行扫描的扫描模块。
  41. 如权利要求40所述的方法，其特征在于，所述扫描模块在至少部分 不同时刻的扫描路径不同。
  42. 如权利要求41所述的方法，其特征在于，所述扫描模块在测距装置的视场内的扫描区域随着时间的累积而增加。
  43. 如权利要求40至42任一项所述的方法，其特征在于，所述扫描模块包括至少一个具有非平行的出射面和入射面的光折射元件。
  44. 如权利要求43所述的方法，其特征在于，所述扫描模块包括在所述光脉冲序列的出射光路上依次排布的2个或3个所述光折射元件。
  45. 如权利要求43所述的方法，其特征在于，所述扫描模块中的至少2个所述光折射元件在扫描过程中旋转，以改变所述光脉冲序列的方向。
  46. 一种点云重采样装置，其特征在于，所述点云重采样装置包括：
    重采样模块，用于对测距装置获取的初始点云进行重采样，以获得重采样后的重采样点云，其中，所述测距装置在扫描视场中具有非均匀的扫描密度，所述重采样点云的分布均匀性高于所述初始点云的分布均匀性，和/或，所述重采样点云的噪声低于所述初始点云的噪声。
  47. 如权利要求46所述的点云重采样装置，其特征在于，所述重采样模块包括：
    插值模块，用于在所述初始点云的点云间隙处进行插值。
  48. 如权利要求47所述的点云重采样装置，其特征在于，所述重采样点云在参考面上的投影点相比所述初始点云在该参考面上的投影点的均匀度高。
  49. 如权利要求48所述的点云重采样装置，其特征在于，所述参考面包括平面、球面、柱面中的至少一种。
  50. 如权利要求49所述的点云重采样装置，其特征在于，
    所述平面为与所述测距装置所发射的光脉冲序列的中心轴相垂直的面；或者，
    所述球面为以所述测距装置为中心的球面；或者，
    所述柱面的中轴线为经过所述测距装置的竖直线。
  51. 如权利要求48所述的点云重采样装置，其特征在于，所述重采样点云在所述参考面上的至少部分投影点均匀分布。
  52. 如权利要求51所述的点云重采样装置，其特征在于，所述重采样点云在所述参考面上的投影点中的至少部分投影点在该参考面的至少一条轨迹 上等间隔排布或者等角度排布。
  53. 如权利要求52所述的点云重采样装置，其特征在于，所述轨迹包括呈规则的直线状、规则的曲线状或者规则的折线状的轨迹。
  54. 如权利要求52所述的点云重采样装置，其特征在于，所述轨迹包括沿水平方向和/或竖直方向上延伸的轨迹。
  55. 如权利要求52所述的点云重采样装置，其特征在于，所述轨迹包括多个同心环。
  56. 如权利要求55所述的点云重采样装置，其特征在于，所述多个同心环的中心位于所述测距装置的中心轴上。
  57. 如权利要求52所述的点云重采样装置，其特征在于，所述轨迹包括一条呈螺旋线的轨迹。
  58. 如权利要求52所述的点云重采样装置，其特征在于，所述轨迹包括多条沿一个中心点往外辐射的多条轨迹。
  59. 如权利要求48所述的点云重采样装置，其特征在于，所述重采样点云相比所述初始点云增加了多个位于同一个所述参考面上的点云点。
  60. 如权利要求48所述的点云重采样装置，其特征在于，所述点云重采样装置还包括图像生成模块，所述图像生成模块包括：
    投影模块，用于将所述初始点云投影到参考面上，得到所述初始点云的投影点；
    所述插值模块，用于在所述参考面上的所述投影点的间隙处确定插值点，以获得位于所述参考面上的图像。
  61. 如权利要求60所述的点云重采样装置，其特征在于，所述插值点的深度值是基于邻近的所述投影点的深度值确定的。
  62. 如权利要求60所述的点云重采样装置，其特征在于，所述图像生成模块还包括离散化模块，用于在确定所述插值点之前在所述参考面上将所述初始点云的投影点进行离散化。
  63. 如权利要求62所述的点云重采样装置，其特征在于，离散化模块具体用于：
    在所述参考面上将所述初始点云的投影点进行网格化，其中，所述插值点的位置包括至少部分空网格。
  64. 如权利要求60-63任一项所述的点云重采样装置，其特征在于，所述重采样模块具体用于：基于所述图像进行重采样，以获得具有特定排布模式的所述重采样点云。
  65. 如权利要求63所述的点云重采样装置，其特征在于，离散化模块具体用于：
    将所述网格的深度值取代该网格内的多个投影点的深度值，其中该网格的深度值是根据落入到该网格中的多个投影点的深度值确定的；以及
    所述插值点的深度值是基于邻近网格的深度值确定的。
  66. 如权利要求63所述的点云重采样装置，其特征在于，所述插值模块还用于：
    确定所述空网格的位置是否天空位置，其中，待插值的所述空网格包括非天空位置的空网格。
  67. 如权利要求63所述的点云重采样装置，其特征在于，所述网格是由所述参考面上多条相互垂直的轨迹形成的。
  68. 如权利要求67所述的点云重采样装置，其特征在于，所述网格是由参考面上多条沿水平方向延伸和多条沿竖直方向延伸的交叉轨迹形成的。
  69. 如权利要求63所述的点云重采样装置，其特征在于，所述网格是由参考面上的多条同心圆环与从该同心环的中心向外辐射的多条轨迹交叉形成的。
  70. 如权利要求63所述的点云重采样装置，其特征在于，所述网格是由参考面上的一条螺旋线，与从该螺旋线的中心向外辐射的多条轨迹交叉形成的。
  71. 如权利要求61至63、65至70任一项所述的点云重采样装置，其特征在于，所述重采样模块还用于：
    在所述初始点云中增加确定的插值点。
  72. 如权利要求46所述的点云重采样装置，其特征在于，所述初始点云不同区域的点云密度不同，所述重采样模块还包括：
    降采样模块，用于在点云密度高的区域对所述初始点云进行降采样。
  73. 如权利要求72所述的点云重采样装置，其特征在于，所述测距装置在扫描视场的中心区域的扫描密度高于其余区域的扫描密度，所述降采样模 块具体用于：对中心区域的初始点云进行降采样。
  74. 如权利要求73所述的点云重采样装置，其特征在于，所述降采样模块具体用于：
    随机或者均匀舍弃所述中心区域的部分网格对应的全部初始点云点，其中，所述网格是由在参考面上将所述初始点云的投影点进行网格化而形成的。
  75. 如权利要求73所述的点云重采样装置，其特征在于，所述降采样模块具体用于：
    限制落入所述中心区域内的至少部分网格内的初始点云点的数量低于阈值数量，其中，所述网格是由在参考面上将初始点云的投影点进行网格化而形成的。
  76. 如权利要求75所述的点云重采样装置，其特征在于，所述降采样模块具体还用于：
    随机或者均匀舍弃所述至少部分网格内的部分初始点云点。
  77. 如权利要求45所述的点云重采样装置，其特征在于，所述重采样模块还包括降噪重采样模块，用于对所述初始点云进行降噪重采样，以使所述重采样点云的噪声低于所述初始点云的噪声。
  78. 如权利要求77所述的点云重采样装置，其特征在于，所述降噪重采样模块具体用于：
    对所述初始点云在参考面上的图像进行滤噪处理。
  79. 如权利要求78所述的点云重采样装置，其特征在于，所述降噪重采样模块具体还用于：
    遍历所述初始点云的每个初始点云点；
    将每个所述初始点云点映射至所述滤噪处理后的所述图像；
    将每个所述初始点云点的深度值调整为其对应像素的深度值，或者，调整为所述初始点云点在所述图像的投影位置的相邻像素的深度值。
  80. 如权利要求79所述的点云重采样装置，其特征在于，所述降噪重采样模块进一步用于：
    基于所述初始点云点调整后的深度值与该初始点云点的初始深度值进行阈值处理，在所述初始深度值和所述调整后的深度值之差大于目标阈值时，舍弃该初始点云点。
  81. 如权利要求77至80任一项所述的点云重采样装置，其特征在于，所述图像包括深度分布图。
  82. 如权利要求81所述的点云重采样装置，其特征在于，还包括用于获得所述图像的图像生成模块，所述图像生成模块包括：
    离散化模块，用于在所述参考面上将所述初始点云的投影点进行网格化，其中，每个网格对应一个像素；
    插值模块，用于对部分空网格进行插值。
  83. 如权利要求81所述的点云重采样装置，其特征在于，所述离散化模块还用于根据落入到网格中的多个投影点的深度值确定该网格的深度值；以及
    所述插值模块，用于基于邻近网格的深度值确定空网格的深度值。
  84. 如权利要求83所述的点云重采样装置，其特征在于，所述插值模块具体用于：
    确定所述空网格的位置是否天空位置，其中，基于邻近网格的深度值确定非天空位置的所述空网格的深度值。
  85. 如权利要求45所述的点云重采样装置，其特征在于，所述测距装置包括用于出射光脉冲序列的发射模块，以及用于改变光脉冲序列的方向以对视场进行扫描的扫描模块。
  86. 如权利要求85所述的点云重采样装置，其特征在于，所述扫描模块在至少部分不同时刻的扫描路径不同。
  87. 如权利要求86所述的点云重采样装置，其特征在于，所述扫描模块在测距装置的视场内的扫描区域随着时间的累积而增加。
  88. 如权利要求85至87任一项所述的点云重采样装置，其特征在于，所述扫描模块包括至少一个具有非平行的出射面和入射面的光折射元件。
  89. 如权利要求88所述的点云重采样装置，其特征在于，所述扫描模块包括在所述光脉冲序列的出射光路上依次排布的2个或3个所述光折射元件。
  90. 如权利要求88所述的点云重采样装置，其特征在于，所述扫描模块中的至少2个所述光折射元件在扫描过程中旋转，以改变所述光脉冲序列的方向。
  91. 一种点云重采样系统，其特征在于，点云重采样系统包括：
    至少一个测距装置，用于探测目标场景生成初始点云；
    存储器，用于存储可执行指令；
    处理器，用于执行所述存储器中存储的所述指令，使得所述处理器执行权利要求1至45中任一项所述的点云重采样的方法。
PCT/CN2019/074585 2019-02-02 2019-02-02 一种点云重采样的方法、装置和系统 WO2020155142A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201980005631.3A CN111801707A (zh) 2019-02-02 2019-02-02 一种点云重采样的方法、装置和系统
PCT/CN2019/074585 WO2020155142A1 (zh) 2019-02-02 2019-02-02 一种点云重采样的方法、装置和系统

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2019/074585 WO2020155142A1 (zh) 2019-02-02 2019-02-02 一种点云重采样的方法、装置和系统

Publications (1)

Publication Number Publication Date
WO2020155142A1 true WO2020155142A1 (zh) 2020-08-06

Family

ID=71839909

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/074585 WO2020155142A1 (zh) 2019-02-02 2019-02-02 一种点云重采样的方法、装置和系统

Country Status (2)

Country Link
CN (1) CN111801707A (zh)
WO (1) WO2020155142A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2023187785A1 (en) * 2022-03-30 2023-10-05 Eeg-Sense Ltd. Method and system for aligning neurophysiological sensors

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101650838A (zh) * 2009-09-04 2010-02-17 浙江工业大学 基于重采样方法和仿射聚类算法的点云简化处理方法
CN102930597A (zh) * 2012-09-28 2013-02-13 中国科学院自动化研究所 一种外存三维模型的处理方法
CN103729872A (zh) * 2013-12-30 2014-04-16 浙江大学 一种基于分段重采样和表面三角化的点云增强方法
US20150006117A1 (en) * 2013-07-01 2015-01-01 Here Global B.V. Learning Synthetic Models for Roof Style Classification Using Point Clouds
CN104268934A (zh) * 2014-09-18 2015-01-07 中国科学技术大学 一种由点云直接重建三维曲面的方法
CN106796728A (zh) * 2016-11-16 2017-05-31 深圳市大疆创新科技有限公司 生成三维点云的方法、装置、计算机系统和移动设备
CN106886980A (zh) * 2015-12-11 2017-06-23 北京智行者科技有限公司 一种基于三维激光雷达目标识别的点云密度增强的方法
WO2018201180A1 (en) * 2017-05-02 2018-11-08 PETRA Data Science Pty Ltd Automated, real time processing, analysis, mapping and reporting of data for the detection of geotechnical features

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101650838A (zh) * 2009-09-04 2010-02-17 浙江工业大学 基于重采样方法和仿射聚类算法的点云简化处理方法
CN102930597A (zh) * 2012-09-28 2013-02-13 中国科学院自动化研究所 一种外存三维模型的处理方法
US20150006117A1 (en) * 2013-07-01 2015-01-01 Here Global B.V. Learning Synthetic Models for Roof Style Classification Using Point Clouds
CN103729872A (zh) * 2013-12-30 2014-04-16 浙江大学 一种基于分段重采样和表面三角化的点云增强方法
CN104268934A (zh) * 2014-09-18 2015-01-07 中国科学技术大学 一种由点云直接重建三维曲面的方法
CN106886980A (zh) * 2015-12-11 2017-06-23 北京智行者科技有限公司 一种基于三维激光雷达目标识别的点云密度增强的方法
CN106796728A (zh) * 2016-11-16 2017-05-31 深圳市大疆创新科技有限公司 生成三维点云的方法、装置、计算机系统和移动设备
WO2018201180A1 (en) * 2017-05-02 2018-11-08 PETRA Data Science Pty Ltd Automated, real time processing, analysis, mapping and reporting of data for the detection of geotechnical features

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
YONGHUI CHEN: "Research on Data Processing Techniques Based on Laser Scanning 3D Point Clouds", DOCTORAL DISSERTATION, no. 02, 10 September 2017 (2017-09-10), CN, pages 1 - 124, XP009522522 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2023187785A1 (en) * 2022-03-30 2023-10-05 Eeg-Sense Ltd. Method and system for aligning neurophysiological sensors

Also Published As

Publication number Publication date
CN111801707A (zh) 2020-10-20

Similar Documents

Publication Publication Date Title
WO2022126427A1 (zh) 点云处理方法、点云处理装置、可移动平台和计算机存储介质
WO2020243962A1 (zh) 物体检测方法、电子设备和可移动平台
WO2021051281A1 (zh) 点云滤噪的方法、测距装置、系统、存储介质和移动平台
US10877622B2 (en) Detection of structured light for depth sensing
WO2020243874A1 (zh) 点云数据的位置坐标的编解码方法、系统和存储介质
WO2022198637A1 (zh) 点云滤噪方法、系统和可移动平台
US20210293928A1 (en) Ranging apparatus, balance method of scan field thereof, and mobile platform
US20210256740A1 (en) Method for increasing point cloud sampling density, point cloud processing system, and readable storage medium
WO2021062581A1 (zh) 路面标识识别方法及装置
WO2022179207A1 (zh) 视窗遮挡检测方法及装置
WO2020072084A1 (en) Depth sensing using grid light patterns
WO2020215252A1 (zh) 测距装置点云滤噪的方法、测距装置和移动平台
JP6919764B2 (ja) レーダ画像処理装置、レーダ画像処理方法、および、プログラム
US11828878B2 (en) LiDAR system and operating method thereof
WO2021232227A1 (zh) 构建点云帧的方法、目标检测方法、测距装置、可移动平台和存储介质
WO2021232247A1 (zh) 点云着色方法、点云着色系统和计算机存储介质
WO2020155142A1 (zh) 一种点云重采样的方法、装置和系统
WO2020237663A1 (zh) 一种多通道激光雷达点云插值的方法和测距装置
WO2020177076A1 (zh) 一种探测装置初始状态标定方法及装置
WO2020087376A1 (zh) 光探测方法、光探测装置和移动平台
US20220082665A1 (en) Ranging apparatus and method for controlling scanning field of view thereof
WO2023004792A1 (zh) 一种激光雷达的姿态标定方法、相关装置、以及存储介质
WO2022256976A1 (zh) 稠密点云真值数据的构建方法、系统和电子设备
WO2022217520A1 (zh) 探测方法、装置、可移动平台及存储介质
WO2020107379A1 (zh) 应用于测距装置的反射率校正方法、测距装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19913068

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19913068

Country of ref document: EP

Kind code of ref document: A1